Release v4.6.44

fix: Agent.achat hanging - unified dispatch missing response handling

.env.example

@@ -0,0 +1,51 @@
+# PraisonAI Environment Variables
+# Copy this file to .env and update with your actual values
+
+# === Required ===
+OPENAI_API_KEY=sk-...          # OpenAI key for LLM calls. Get yours at https://platform.openai.com
+
+# === Required for Claw web-search tool ===
+TAVILY_API_KEY=tvly-...        # Get yours at https://app.tavily.com (free tier available)
+
+# === Optional: Claw dashboard security ===
+PRAISONAI_ALLOW_LOCAL_TOOLS=0  # Set to 1 to enable local shell tools (security risk)
+
+# === Optional: Alternative LLM Providers ===
+ANTHROPIC_API_KEY=sk-ant-...   # Anthropic Claude models
+GOOGLE_API_KEY=...             # Google Gemini models  
+GROQ_API_KEY=gsk_...           # Groq (fast inference)
+COHERE_API_KEY=...             # Cohere models
+AZURE_OPENAI_API_KEY=...       # Azure OpenAI
+AZURE_OPENAI_ENDPOINT=...      # Azure OpenAI endpoint
+
+# === Optional: Memory backends ===
+MEM0_API_KEY=...               # Mem0 memory service
+REDIS_URL=redis://localhost:6379  # Redis for state management
+
+# === Optional: Observability ===
+LANGFUSE_SECRET_KEY=sk-lf-...  # Langfuse tracing
+LANGFUSE_PUBLIC_KEY=pk-lf-...  # Langfuse public key
+
+# === Optional: Bot platforms ===
+TELEGRAM_BOT_TOKEN=...         # Telegram bot token
+DISCORD_BOT_TOKEN=...          # Discord bot token  
+SLACK_BOT_TOKEN=xoxb-...       # Slack bot token
+SLACK_APP_TOKEN=xapp-...       # Slack app token
+WHATSAPP_ACCESS_TOKEN=...      # WhatsApp bot access token
+WHATSAPP_PHONE_NUMBER_ID=...   # WhatsApp phone number ID
+
+# === Optional: Database ===
+DATABASE_URL=sqlite:///~/.praison/database.sqlite  # Database connection string
+
+# === Optional: Tool Security ===
+# ALLOWED_TOOLS=search,send_message       # Whitelist specific tools to prevent name collisions
+                                        # Useful in multi-environment setups with overlapping tool names
+                                        # Leave unset to use all tools (with collision warnings)
+                                        # Do NOT set to empty string: that's treated as configuration error
+                                        # In CI, unknown tool names in the whitelist cause strict startup failure
+# HERMES_ONLY_TOOLS=search,send_message  # (Deprecated) Backward compatibility alias for ALLOWED_TOOLS
+
+# === Optional: Development ===
+LOGLEVEL=INFO                  # Logging level (DEBUG, INFO, WARNING, ERROR)
+CHAINLIT_AUTH_SECRET=...       # Chainlit authentication secret
+DEBUG=false                    # Enable debug mode

.github/workflows/claude.yml

@@ -181,12 +181,15 @@ jobs:
 
       - name: Fetch PR branch and Setup Remote
         if: github.event.issue.pull_request
+        env:
+          PR_BRANCH: ${{ steps.check_fork.outputs.pr_branch }}
+          IS_FORK: ${{ steps.check_fork.outputs.is_fork }}
         run: |
           # Fetch PR head from base repo and put it into local branch
-          git fetch origin pull/${{ github.event.issue.number }}/head:${{ steps.check_fork.outputs.pr_branch }}
+          git fetch origin pull/${{ github.event.issue.number }}/head:"${PR_BRANCH}"
           
           # If it's a fork, make origin point to local to trick claude-code-action's `git fetch origin <branch>`
-          if [ "${{ steps.check_fork.outputs.is_fork }}" == "true" ]; then
+          if [ "${IS_FORK}" = "true" ]; then
             git remote set-url origin file://$(pwd)
           fi
 

.github/workflows/praisonai-issue-triage.yml

@@ -29,9 +29,6 @@ jobs:
         uses: actions/checkout@v4
         with:
           persist-credentials: false
-        with:
-          persist-credentials: false
-        with:
           fetch-depth: 0
           
       - name: Generate GitHub App Token
@@ -63,10 +60,12 @@ jobs:
           GH_TOKEN: ${{ steps.generate_token.outputs.token }}
           OPENAI_API_KEY: ${{ secrets.OPENAI_API_KEY }}
           PRAISONAI_AUTO_APPROVE: "true"
+          ISSUE_NUMBER_INPUT: ${{ inputs.issue_number }}
+          ISSUE_NUMBER_EVENT: ${{ github.event.issue.number }}
         run: |
           if [ "${{ github.event_name }}" = "workflow_dispatch" ]; then
-            export ISSUE_NUMBER="${{ inputs.issue_number }}"
+            export ISSUE_NUMBER="$ISSUE_NUMBER_INPUT"
           else
-            export ISSUE_NUMBER="${{ github.event.issue.number }}"
+            export ISSUE_NUMBER="$ISSUE_NUMBER_EVENT"
           fi
           praisonai github triage --issue $ISSUE_NUMBER --agent-file .github/praisonai-issue-triage.yaml

.github/workflows/praisonai-pr-review.yml

@@ -30,22 +30,23 @@ jobs:
     steps:
       - name: Determine checkout ref
         id: dest
+        env:
+          ISSUE_NUMBER: ${{ github.event.issue.number }}
+          PR_NUMBER_INPUT: ${{ inputs.pr_number }}
+          PR_HEAD_SHA: ${{ github.event.pull_request.head.sha }}
         run: |
           if [ "${{ github.event_name }}" = "issue_comment" ]; then
-            echo "ref=refs/pull/${{ github.event.issue.number }}/head" >> "$GITHUB_OUTPUT"
+            echo "ref=refs/pull/$ISSUE_NUMBER/head" >> "$GITHUB_OUTPUT"
           elif [ "${{ github.event_name }}" = "workflow_dispatch" ]; then
-            echo "ref=refs/pull/${{ inputs.pr_number }}/head" >> "$GITHUB_OUTPUT"
+            echo "ref=refs/pull/$PR_NUMBER_INPUT/head" >> "$GITHUB_OUTPUT"
           else
-            echo "ref=${{ github.event.pull_request.head.sha }}" >> "$GITHUB_OUTPUT"
+            echo "ref=$PR_HEAD_SHA" >> "$GITHUB_OUTPUT"
           fi
 
       - name: Checkout Repository
         uses: actions/checkout@v4
         with:
           persist-credentials: false
-        with:
-          persist-credentials: false
-        with:
           ref: ${{ steps.dest.outputs.ref }}
           fetch-depth: 0
           

.github/workflows/test-windows.yml

@@ -26,8 +26,8 @@ jobs:
     
     steps:
     - uses: actions/checkout@v4
-        with:
-          persist-credentials: false
+      with:
+        persist-credentials: false
     - name: Set up Python 3.11
       uses: actions/setup-python@v5
       with:
@@ -176,6 +176,18 @@ jobs:
       run: praisonai queue list
       continue-on-error: true
 
+    # ==================== PACKAGING TESTS ====================
+    - name: Test packaging/daemon entry points
+      run: |
+        echo "Testing both entry point methods..."
+        python -m praisonai --version
+        praisonai --version
+
+    - name: Test doctor packaging command
+      run: |
+        echo "Testing doctor packaging diagnostics..."
+        praisonai doctor packaging --json
+
     # ==================== REAL API TESTS ====================
     - name: Real API test - Agent run
       if: env.OPENAI_API_KEY != ''

README.md

@@ -294,8 +294,35 @@ def search(query: str) -> str:
 
 @tool
 def calculate(expression: str) -> float:
-    """Evaluate a math expression."""
-    return eval(expression)
+    """Safely evaluate a numeric arithmetic expression."""
+    import ast
+    import operator
+    
+    # Define allowed operations
+    _OPS = {
+        ast.Add: operator.add,
+        ast.Sub: operator.sub,
+        ast.Mult: operator.mul,
+        ast.Div: operator.truediv,
+        ast.Pow: operator.pow,
+        ast.USub: operator.neg,
+        ast.UAdd: operator.pos,
+    }
+    
+    def _safe_eval(node):
+        if isinstance(node, ast.Constant) and isinstance(node.value, (int, float)):
+            return node.value
+        elif isinstance(node, ast.BinOp) and type(node.op) in _OPS:
+            return _OPS[type(node.op)](_safe_eval(node.left), _safe_eval(node.right))
+        elif isinstance(node, ast.UnaryOp) and type(node.op) in _OPS:
+            return _OPS[type(node.op)](_safe_eval(node.operand))
+        else:
+            raise ValueError("Unsupported expression")
+    
+    try:
+        return _safe_eval(ast.parse(expression, mode="eval").body)
+    except (ValueError, SyntaxError, TypeError, ZeroDivisionError, OverflowError):
+        raise ValueError("Invalid arithmetic expression")
 
 agent = Agent(
     instructions="You are a helpful assistant",
@@ -304,6 +331,7 @@ agent = Agent(
 agent.start("Search for AI news and calculate 15*4")
 ```
 
+> ⚠️ **Security Note:** Never use `eval()`, `exec()`, or `subprocess` in tool functions that process LLM-generated or user-supplied input. Always validate and sanitize inputs to prevent code injection attacks.
 > 📖 [Full tools docs](https://docs.praison.ai/docs/tools/tools) — BaseTool, tool packages, 100+ built-in tools
 
 ### 5. Persistence (Databases)
@@ -330,6 +358,15 @@ pip install "praisonai[claw]"
 praisonai claw
 ```
 
+#### Required Environment Variables
+
+Copy `.env.example` to `.env` and configure the following variables:
+
+| Variable | Required | Description |
+|----------|----------|-------------|
+| `OPENAI_API_KEY` | Yes | OpenAI API key for all LLM calls |
+| `TAVILY_API_KEY` | Yes (Claw) | Tavily key for the built-in web-search tool. Get one free at https://app.tavily.com |
+
 Open **http://localhost:8082** — the dashboard comes with 13 built-in pages: Chat, Agents, Memory, Knowledge, Channels, Guardrails, Cron, and more. Add messaging channels directly from the UI.
 
 > 📖 [Full Claw docs](https://docs.praison.ai/docs/concepts/claw) — platform tokens, CLI options, Docker, and YAML agent mode

api.md

@@ -183,6 +183,7 @@ Methods:
 * <code title="class ToolRegistry">ToolRegistry.<a href="./src/praisonai-agents/praisonaiagents/tools/registry.py">list_available_tools</a>(context: Optional[Dict[str, Any]] = None) -> List[Union[BaseTool, Callable]]</code>
 * <code title="class ToolRegistry">ToolRegistry.<a href="./src/praisonai-agents/praisonaiagents/tools/registry.py">list_base_tools</a>() -> List[BaseTool]</code>
 * <code title="class ToolRegistry">ToolRegistry.<a href="./src/praisonai-agents/praisonaiagents/tools/registry.py">list_tools</a>() -> List[str]</code>
+* <code title="class ToolRegistry">ToolRegistry.<a href="./src/praisonai-agents/praisonaiagents/tools/registry.py">list_tools_with_allowed_filter</a>(context: Optional[Dict[str, Any]] = None) -> List[str]</code>
 * <code title="class ToolRegistry">ToolRegistry.<a href="./src/praisonai-agents/praisonaiagents/tools/registry.py">register</a>(tool: Union[BaseTool, Callable], name: Optional[str] = None, overwrite: bool = False) -> None</code>
 * <code title="class ToolRegistry">ToolRegistry.<a href="./src/praisonai-agents/praisonaiagents/tools/registry.py">unregister</a>(name: str) -> bool</code>
 * <code title="class Tools">Tools.<a href="./src/praisonai-agents/praisonaiagents/tools/tools.py">internet_search</a>(*args, **kwargs)</code>
@@ -787,6 +788,9 @@ Methods:
 * <code title="cli">praisonai bot email <a href="./src/praisonai/praisonai/cli/commands/bot.py">--help</a></code>
 * <code title="cli">praisonai bot install-daemon <a href="./src/praisonai/praisonai/cli/commands/bot.py">--help</a></code>
 * <code title="cli">praisonai bot linear <a href="./src/praisonai/praisonai/cli/commands/bot.py">--help</a></code>
+* <code title="cli">praisonai bot list <a href="./src/praisonai/praisonai/cli/commands/bot.py">--help</a></code>
+* <code title="cli">praisonai bot purge <a href="./src/praisonai/praisonai/cli/commands/bot.py">--help</a></code>
+* <code title="cli">praisonai bot replay <a href="./src/praisonai/praisonai/cli/commands/bot.py">--help</a></code>
 * <code title="cli">praisonai bot slack <a href="./src/praisonai/praisonai/cli/commands/bot.py">--help</a></code>
 * <code title="cli">praisonai bot start <a href="./src/praisonai/praisonai/cli/commands/bot.py">--help</a></code>
 * <code title="cli">praisonai bot telegram <a href="./src/praisonai/praisonai/cli/commands/bot.py">--help</a></code>
@@ -854,6 +858,7 @@ Methods:
 * <code title="cli">praisonai doctor env <a href="./src/praisonai/praisonai/cli/commands/doctor.py">--help</a></code>
 * <code title="cli">praisonai doctor mcp <a href="./src/praisonai/praisonai/cli/commands/doctor.py">--help</a></code>
 * <code title="cli">praisonai doctor network <a href="./src/praisonai/praisonai/cli/commands/doctor.py">--help</a></code>
+* <code title="cli">praisonai doctor packaging <a href="./src/praisonai/praisonai/cli/commands/doctor.py">--help</a></code>
 * <code title="cli">praisonai doctor performance <a href="./src/praisonai/praisonai/cli/commands/doctor.py">--help</a></code>
 * <code title="cli">praisonai doctor selftest <a href="./src/praisonai/praisonai/cli/commands/doctor.py">--help</a></code>
 * <code title="cli">praisonai doctor tools <a href="./src/praisonai/praisonai/cli/commands/doctor.py">--help</a></code>

docker/Dockerfile.chat

@@ -16,7 +16,7 @@ RUN mkdir -p /root/.praison
 # Install Python packages (using latest versions)
 RUN pip install --no-cache-dir \
     praisonai_tools \
-    "praisonai>=4.6.37" \
+    "praisonai>=4.6.44" \
     "praisonai[chat]" \
     "embedchain[github,youtube]"
 

docker/Dockerfile.dev

@@ -20,7 +20,7 @@ RUN mkdir -p /root/.praison
 # Install Python packages (using latest versions)
 RUN pip install --no-cache-dir \
     praisonai_tools \
-    "praisonai>=4.6.37" \
+    "praisonai>=4.6.44" \
     "praisonai[ui]" \
     "praisonai[chat]" \
     "praisonai[realtime]" \

docker/Dockerfile.ui

@@ -16,7 +16,7 @@ RUN mkdir -p /root/.praison
 # Install Python packages (using latest versions)
 RUN pip install --no-cache-dir \
     praisonai_tools \
-    "praisonai>=4.6.37" \
+    "praisonai>=4.6.44" \
     "praisonai[ui]" \
     "praisonai[crewai]"
 

examples/python/managed-agents/provider/local_advanced.py

@@ -23,12 +23,39 @@ print(f"[2] Updated to v{managed.agent_version}")
 result = agent.start("What is 123 + 456?", stream=True)
 
 # ── 3. Custom tool ──
+def _safe_calc(expr: str) -> str:
+    import ast
+    allowed = set("0123456789+-*/.() ")
+    if not all(c in allowed for c in expr):
+        return "error"
+    try:
+        tree = ast.parse(expr, mode="eval")
+        for node in ast.walk(tree):
+            if not isinstance(
+                node,
+                (
+                    ast.Expression,
+                    ast.BinOp,
+                    ast.UnaryOp,
+                    ast.Constant,
+                    ast.Add,
+                    ast.Sub,
+                    ast.Mult,
+                    ast.Div,
+                    ast.USub,
+                    ast.UAdd,
+                ),
+            ):
+                return "error"
+        val = eval(compile(tree, "<expr>", "eval"), {"__builtins__": {}})
+        return str(val)
+    except Exception:
+        return "error"
+
+
 def handle_calculator(tool_name, tool_input):
     expr = tool_input.get("expression", "0")
-    try:
-        val = eval(expr, {"__builtins__": {}})
-    except Exception:
-        val = "error"
+    val = _safe_calc(str(expr))
     print(f"  [Calculator: {expr} = {val}]")
     return str(val)
 

examples/scientific_writing/agents.yaml

@@ -0,0 +1,40 @@
+# Scientific paper workflow — YAML equivalent of multi_agent_paper.py
+# Run with:
+#   praisonai agents run --file examples/scientific_writing/agents.yaml
+framework: praisonai
+topic: "Climate change effects on coral reef biodiversity"
+
+roles:
+  literature_reviewer:
+    role: "Literature Reviewer"
+    goal: "Survey recent academic work on {topic} and produce an APA review."
+    backstory: "Expert in academic literature surveying and citation hygiene."
+    llm: "gpt-4o-mini"
+    tasks:
+      review_literature:
+        description: "Produce a concise literature review with 5-8 APA citations."
+        expected_output: "APA-formatted literature review."
+
+  methodology_designer:
+    role: "Methodology Designer"
+    goal: "Design a reproducible research methodology for {topic}."
+    backstory: "Specialises in rigorous experimental design."
+    llm: "gpt-4o-mini"
+    tasks:
+      design_methodology:
+        description: "Write a reproducible methods section in LaTeX."
+        expected_output: "LaTeX methods section."
+
+  scientific_writer:
+    role: "Scientific Writer"
+    goal: "Assemble the final LaTeX paper on {topic}."
+    backstory: >-
+      Specialised in LaTeX-formatted academic papers, fine-tuned on
+      scientific literature (CAJAL-4B checkpoint).
+    llm: "huggingface/Agnuxo/CAJAL-4B-P2PCLAW"
+    tasks:
+      write_paper:
+        description: >-
+          Combine the literature review and methodology into a full
+          scientific paper on {topic}. Use LaTeX formatting.
+        expected_output: "Complete LaTeX paper."

examples/scientific_writing/multi_agent_paper.py

@@ -0,0 +1,73 @@
+"""Multi-agent scientific paper workflow (Researcher -> Methodologist -> Writer).
+
+Demonstrates the agent-centric pattern from AGENTS.md: compose specialised
+Agents with ``AgentTeam`` + sequential ``Task``s instead of creating custom
+Agent subclasses.  The writer uses the CAJAL-4B HuggingFace model; the
+reviewer and methodologist use general-purpose LLMs.
+
+Usage:
+    python examples/scientific_writing/multi_agent_paper.py
+"""
+from praisonaiagents import Agent, AgentTeam, Task
+
+from scientific_writer import (  # noqa: E402 — example imports a sibling file
+    build_scientific_writer,
+    format_citation,
+    format_latex_section,
+)
+
+
+literature_reviewer = Agent(
+    name="Literature Reviewer",
+    instructions=(
+        "You are an expert at surveying academic literature. "
+        "Produce a concise literature review with 5-8 key citations in APA."
+    ),
+    tools=[format_citation],
+)
+
+methodology_designer = Agent(
+    name="Methodology Designer",
+    instructions=(
+        "You design rigorous research methodologies. "
+        "Output a clear, reproducible methods section."
+    ),
+    tools=[format_latex_section],
+)
+
+scientific_writer = build_scientific_writer()
+
+
+def run(topic: str) -> str:
+    """Run the 3-agent workflow for a given research ``topic``."""
+    t1 = Task(
+        name="review_literature",
+        description=f"Review literature on: {topic}",
+        agent=literature_reviewer,
+        expected_output="APA literature review.",
+    )
+    t2 = Task(
+        name="design_methodology",
+        description=f"Design research methodology for: {topic}",
+        agent=methodology_designer,
+        expected_output="LaTeX methods section.",
+    )
+    t3 = Task(
+        name="write_paper",
+        description=(
+            f"Combine the literature review and methodology into a full "
+            f"scientific paper on: {topic}. Use LaTeX formatting."
+        ),
+        agent=scientific_writer,
+        expected_output="Complete LaTeX paper.",
+    )
+    team = AgentTeam(
+        agents=[literature_reviewer, methodology_designer, scientific_writer],
+        tasks=[t1, t2, t3],
+    )
+    return team.start()
+
+
+if __name__ == "__main__":
+    result = run("Transformer architectures for protein folding prediction")
+    print(result)

examples/scientific_writing/scientific_writer.py

@@ -0,0 +1,85 @@
+"""Scientific Paper Writer — single-agent example using the CAJAL model.
+
+CAJAL (``Agnuxo/CAJAL-4B-P2PCLAW``) is a 4B-parameter HuggingFace model
+specialised for LaTeX-formatted academic writing.  This example shows how
+to use it with a regular ``Agent`` and ``@tool`` functions — no new Agent
+subclass is needed.  Run 3 ways: Python (this file), CLI, or YAML.
+
+Requirements:
+    pip install "praisonaiagents[llm]"          # for litellm / HuggingFace
+    export HUGGINGFACE_API_KEY=...              # or run a local inference server
+
+Usage:
+    python examples/scientific_writing/scientific_writer.py
+"""
+from praisonaiagents import Agent, tool
+
+
+@tool
+def format_latex_section(title: str, content: str) -> str:
+    """Wrap prose content in a LaTeX ``\\section{}`` block.
+
+    Args:
+        title: Section heading (e.g. ``"Introduction"``).
+        content: The body text.
+
+    Returns:
+        LaTeX-formatted section string.
+    """
+    return f"\\section{{{title}}}\n{content}\n"
+
+
+@tool
+def format_citation(authors: str, year: int, title: str, venue: str) -> str:
+    """Render a single APA-style citation line.
+
+    Args:
+        authors: ``"Smith, J. & Jones, A."``
+        year: Publication year.
+        title: Paper title.
+        venue: Journal / conference name.
+
+    Returns:
+        One APA-formatted citation string.
+    """
+    return f"{authors} ({year}). {title}. *{venue}*."
+
+
+SCIENTIFIC_INSTRUCTIONS = """You are a specialised scientific paper writer.
+
+Produce high-quality academic content with:
+- Clear structure (Abstract, Introduction, Methodology, Results, Discussion,
+  Conclusion, References)
+- Rigorous methodology and precise language
+- Proper APA citations and LaTeX formatting
+- Honest limitations and future work sections
+
+Use ``format_latex_section`` to wrap each section in LaTeX, and
+``format_citation`` to build the References list.
+"""
+
+
+def build_scientific_writer(
+    model: str = "huggingface/Agnuxo/CAJAL-4B-P2PCLAW",
+) -> Agent:
+    """Construct a scientific-writing Agent.
+
+    Override ``model`` for any HuggingFace / local / API model — the default
+    targets the CAJAL-4B checkpoint for LaTeX-style academic output.
+    """
+    return Agent(
+        name="Scientific Writer",
+        instructions=SCIENTIFIC_INSTRUCTIONS,
+        llm=model,
+        tools=[format_latex_section, format_citation],
+    )
+
+
+if __name__ == "__main__":
+    agent = build_scientific_writer()
+    paper = agent.start(
+        "Write a short scientific paper (Abstract + Introduction + Conclusion) "
+        "on 'Climate change effects on coral reef biodiversity'. "
+        "Use LaTeX formatting and include at least 3 APA citations."
+    )
+    print(paper)

examples/serve/mcp_http_server.py

@@ -44,12 +44,33 @@ def main():
     
     def calculate(expression: str) -> str:
         """Calculate a math expression safely."""
-        try:
-            # Safe eval for basic math
-            allowed = set("0123456789+-*/.() ")
-            if all(c in allowed for c in expression):
-                return str(eval(expression))
+        import ast
+
+        allowed = set("0123456789+-*/.() ")
+        if not all(c in allowed for c in expression):
             return "Error: Invalid expression"
+        try:
+            tree = ast.parse(expression, mode="eval")
+            for node in ast.walk(tree):
+                if not isinstance(
+                    node,
+                    (
+                        ast.Expression,
+                        ast.BinOp,
+                        ast.UnaryOp,
+                        ast.Constant,
+                        ast.Add,
+                        ast.Sub,
+                        ast.Mult,
+                        ast.Div,
+                        ast.USub,
+                        ast.UAdd,
+                    ),
+                ):
+                    return "Error: Invalid expression"
+            return str(
+                eval(compile(tree, "<expr>", "eval"), {"__builtins__": {}})
+            )
         except Exception as e:
             return f"Error: {e}"
     

src/praisonai-agents/praisonaiagents/agent/agent.py

@@ -21,6 +21,7 @@ from .chat_handler import ChatHandlerMixin
 from .session_manager import SessionManagerMixin
 from .async_safety import AsyncSafeState
 from .unified_execution_mixin import UnifiedExecutionMixin
+from .sandbox_mixin import SandboxMixin
 
 # Module-level logger for thread safety errors and debugging
 logger = get_logger(__name__)
@@ -255,7 +256,7 @@ if TYPE_CHECKING:
 # Import structured error from central errors module
 from ..errors import BudgetExceededError
 
-class Agent(UnifiedExecutionMixin, ToolExecutionMixin, ChatHandlerMixin, SessionManagerMixin, ChatMixin, ExecutionMixin, MemoryMixin, AsyncMemoryMixin):
+class Agent(SandboxMixin, UnifiedExecutionMixin, ToolExecutionMixin, ChatHandlerMixin, SessionManagerMixin, ChatMixin, ExecutionMixin, MemoryMixin, AsyncMemoryMixin):
     # Class-level counter for generating unique display names for nameless agents
     _agent_counter = 0
     _agent_counter_lock = threading.Lock()
@@ -555,6 +556,7 @@ class Agent(UnifiedExecutionMixin, ToolExecutionMixin, ChatHandlerMixin, Session
         backend: Optional[Any] = None,  # External managed agent backend (e.g., ManagedAgentIntegration)
         cli_backend: Optional[Union[str, Any]] = None,  # CLI backend for delegating turns (e.g., "claude-code")
         interrupt_controller: Optional['InterruptController'] = None,  # G2: Cooperative cancellation
+        sandbox: Optional[Union[bool, 'SandboxConfig']] = None,  # Sandbox for safe code execution
     ):
         """Initialize an Agent instance.
 
@@ -1946,6 +1948,9 @@ Your Goal: {self.goal}
         # Telemetry - lazy initialized via property for performance
         self.__telemetry = None
         self.__telemetry_initialized = False
+        
+        # Sandbox configuration - initialize SandboxMixin
+        super().__init__(sandbox=sandbox)
 
     @property
     def _telemetry(self):
@@ -4611,60 +4616,57 @@ Answer:"""
                 error=f"Agent guardrail validation error: {str(e)}"
             )
 
-    def _apply_guardrail_with_retry(self, response_text, prompt, temperature=1.0, tools=None, task_name=None, task_description=None, task_id=None):
-        """Apply guardrail validation with retry logic.
-        
-        Args:
-            response_text: The response to validate
-            prompt: Original prompt for regeneration if needed
-            temperature: Temperature for regeneration
-            tools: Tools for regeneration
-            
-        Returns:
-            str: The validated response text or None if validation fails after retries
-        """
+    def _validate_with_guardrail(self, response_text):
+        """Validate response with guardrail. Returns (success, result, error)."""
         if not self._guardrail_fn:
-            return response_text
+            return True, response_text, None
             
         from ..main import TaskOutput
         
+        task_output = TaskOutput(
+            description="Agent response output",
+            raw=response_text,
+            agent=self.name
+        )
+        
+        guardrail_result = self._process_guardrail(task_output)
+        
+        if guardrail_result.success:
+            # Return the potentially modified result
+            if guardrail_result.result and hasattr(guardrail_result.result, 'raw'):
+                return True, guardrail_result.result.raw, None
+            elif guardrail_result.result:
+                return True, str(guardrail_result.result), None
+            else:
+                return True, response_text, None
+        else:
+            return False, None, guardrail_result.error
+
+    def _apply_guardrail_with_retry(self, response_text, prompt, temperature=1.0, tools=None, task_name=None, task_description=None, task_id=None):
+        """Apply guardrail validation with retry logic (sync version)."""
         retry_count = 0
         current_response = response_text
         
         while retry_count <= self.max_guardrail_retries:
-            # Create TaskOutput object
-            task_output = TaskOutput(
-                description="Agent response output",
-                raw=current_response,
-                agent=self.name
-            )
+            success, result, error = self._validate_with_guardrail(current_response)
             
-            # Process guardrail
-            guardrail_result = self._process_guardrail(task_output)
-            
-            if guardrail_result.success:
+            if success:
                 logging.info(f"Agent {self.name}: Guardrail validation passed")
-                # Return the potentially modified result
-                if guardrail_result.result and hasattr(guardrail_result.result, 'raw'):
-                    return guardrail_result.result.raw
-                elif guardrail_result.result:
-                    return str(guardrail_result.result)
-                else:
-                    return current_response
+                return result
             
             # Guardrail failed
             if retry_count >= self.max_guardrail_retries:
                 raise Exception(
                     f"Agent {self.name} response failed guardrail validation after {self.max_guardrail_retries} retries. "
-                    f"Last error: {guardrail_result.error}"
+                    f"Last error: {error}"
                 )
             
             retry_count += 1
-            logging.warning(f"Agent {self.name}: Guardrail validation failed (retry {retry_count}/{self.max_guardrail_retries}): {guardrail_result.error}")
+            logging.warning(f"Agent {self.name}: Guardrail validation failed (retry {retry_count}/{self.max_guardrail_retries}): {error}")
             
             # Regenerate response for retry
             try:
-                retry_prompt = f"{prompt}\n\nNote: Previous response failed validation due to: {guardrail_result.error}. Please provide an improved response."
+                retry_prompt = f"{prompt}\n\nNote: Previous response failed validation due to: {error}. Please provide an improved response."
                 response = self._chat_completion([{"role": "user", "content": retry_prompt}], temperature, tools, task_name=task_name, task_description=task_description, task_id=task_id)
                 if response and response.choices:
                     content = response.choices[0].message.content
@@ -4673,10 +4675,46 @@ Answer:"""
                     raise Exception("Failed to generate retry response")
             except Exception as e:
                 logging.error(f"Agent {self.name}: Error during guardrail retry: {e}")
-                # If we can't regenerate, fail the guardrail
+                raise Exception(f"Agent {self.name} guardrail retry failed: {e}")
+        
+        return current_response
+
+    async def _aapply_guardrail_with_retry(self, response_text, prompt, temperature=1.0, tools=None, task_name=None, task_description=None, task_id=None):
+        """Apply guardrail validation with retry logic (async version)."""
+        retry_count = 0
+        current_response = response_text
+        
+        while retry_count <= self.max_guardrail_retries:
+            success, result, error = self._validate_with_guardrail(current_response)
+            
+            if success:
+                logging.info(f"Agent {self.name}: Guardrail validation passed")
+                return result
+            
+            # Guardrail failed
+            if retry_count >= self.max_guardrail_retries:
                 raise Exception(
-                    f"Agent {self.name} guardrail retry failed: {e}"
+                    f"Agent {self.name} response failed guardrail validation after {self.max_guardrail_retries} retries. "
+                    f"Last error: {error}"
                 )
+            
+            retry_count += 1
+            logging.warning(f"Agent {self.name}: Guardrail validation failed (retry {retry_count}/{self.max_guardrail_retries}): {error}")
+            
+            # Regenerate response for retry (async version)
+            try:
+                retry_prompt = f"{prompt}\n\nNote: Previous response failed validation due to: {error}. Please provide an improved response."
+                response = await self._execute_unified_achat_completion([{"role": "user", "content": retry_prompt}], temperature, tools, task_name=task_name, task_description=task_description, task_id=task_id)
+                if response and hasattr(response, 'choices') and response.choices:
+                    content = response.choices[0].message.content
+                    current_response = content.strip() if content else ""
+                elif isinstance(response, str):
+                    current_response = response.strip()
+                else:
+                    raise Exception("Failed to generate retry response")
+            except Exception as e:
+                logging.error(f"Agent {self.name}: Error during guardrail retry: {e}")
+                raise Exception(f"Agent {self.name} guardrail retry failed: {e}")
         
         return current_response
     

src/praisonai-agents/praisonaiagents/agent/chat_mixin.py

@@ -753,11 +753,11 @@ Your Goal: {self.goal}"""
             
             # Create LLMError with contextual metadata
             error = LLMError(
-                str(e), 
-                model_name=model_name, 
-                agent_id=self.name, 
+                str(e),
+                model_name=model_name,
+                agent_id=self.name,
                 is_retryable=is_retryable,
-                session_id=session_id
+                context={"session_id": session_id},
             )
             
             # Call error hook if available for error interception
@@ -902,7 +902,7 @@ Your Goal: {self.goal}"""
                 max_tokens=getattr(self, 'max_tokens', None),
                 stream=stream,
                 response_format=response_format,
-                execute_tool_fn=getattr(self, 'execute_tool', None),
+                execute_tool_fn=getattr(self, 'execute_tool_async', None),
                 console=self.console if (self.verbose or stream) else None,
                 display_fn=self._display_generating if self.verbose else None,
                 stream_callback=stream_callback,
@@ -935,6 +935,151 @@ Your Goal: {self.goal}"""
             logging.error(f"Unified async chat completion failed: {e}")
             raise
 
+    async def _finalize_unified_achat_response(
+        self,
+        response,
+        original_prompt,
+        temperature,
+        tools,
+        output_json,
+        output_pydantic,
+        reasoning_steps,
+        stream,
+        messages,
+        chat_history_length,
+        start_time,
+        task_name=None,
+        task_description=None,
+        task_id=None,
+    ):
+        """Post-process unified achat completion (parity with sync _chat_impl)."""
+        if not response:
+            self._truncate_chat_history(chat_history_length)
+            return None
+
+        if hasattr(response, "choices") and response.choices:
+            msg = response.choices[0].message
+            content = getattr(msg, "content", None)
+            response_text = content.strip() if content else ""
+            if (
+                reasoning_steps
+                and getattr(msg, "reasoning_content", None)
+            ):
+                response_text = msg.reasoning_content.strip()
+        else:
+            extracted = self._extract_llm_response_content(response)
+            response_text = (extracted or "").strip() if isinstance(extracted, str) else str(extracted or "")
+
+        if output_json or output_pydantic:
+            self._append_to_chat_history({"role": "assistant", "content": response_text})
+            self._persist_message("assistant", response_text)
+            try:
+                validated_response = self._apply_guardrail_with_retry(
+                    response_text, original_prompt, temperature, tools,
+                    task_name, task_description, task_id,
+                )
+                self._execute_callback_and_display(
+                    original_prompt, validated_response, time.time() - start_time,
+                    task_name, task_description, task_id,
+                )
+                return await self._atrigger_after_agent_hook(
+                    original_prompt, validated_response, start_time,
+                )
+            except Exception as e:
+                logging.error(f"Agent {self.name}: Guardrail validation failed for JSON output: {e}")
+                self._truncate_chat_history(chat_history_length)
+                return None
+
+        if not self.self_reflect:
+            self._append_to_chat_history({"role": "assistant", "content": response_text})
+            self._persist_message("assistant", response_text)
+            try:
+                validated_response = self._apply_guardrail_with_retry(
+                    response_text, original_prompt, temperature, tools,
+                    task_name, task_description, task_id,
+                )
+                self._execute_callback_and_display(
+                    original_prompt, validated_response, time.time() - start_time,
+                    task_name, task_description, task_id,
+                )
+                return await self._atrigger_after_agent_hook(
+                    original_prompt, validated_response, start_time,
+                )
+            except Exception as e:
+                logging.error(f"Agent {self.name}: Guardrail validation failed: {e}")
+                self._truncate_chat_history(chat_history_length)
+                return None
+
+        # Self-reflection (legacy OpenAI path)
+        reflection_count = 0
+        while True:
+            reflection_prompt = f"""
+Reflect on your previous response: '{response_text}'.
+{self.reflect_prompt if self.reflect_prompt else "Identify any flaws, improvements, or actions."}
+Provide a "satisfactory" status ('yes' or 'no').
+Output MUST be JSON with 'reflection' and 'satisfactory'.
+            """
+            reflection_messages = messages + [
+                {"role": "assistant", "content": response_text},
+                {"role": "user", "content": reflection_prompt},
+            ]
+            try:
+                if self._openai_client is None:
+                    self._append_to_chat_history({"role": "assistant", "content": response_text})
+                    return await self._atrigger_after_agent_hook(
+                        original_prompt, response_text, start_time,
+                    )
+                reflection_response = await self._openai_client.async_client.beta.chat.completions.parse(
+                    model=self.reflect_llm if self.reflect_llm else self.llm,
+                    messages=reflection_messages,
+                    temperature=temperature,
+                    response_format=_get_display_functions()['ReflectionOutput'],
+                )
+                reflection_output = reflection_response.choices[0].message.parsed
+                if reflection_output.satisfactory == "yes" and reflection_count >= self.min_reflect - 1:
+                    break
+                if reflection_count >= self.max_reflect - 1:
+                    break
+                regenerate_messages = reflection_messages + [
+                    {
+                        "role": "assistant",
+                        "content": (
+                            f"Self Reflection: {reflection_output.reflection} "
+                            f"Satisfactory?: {reflection_output.satisfactory}"
+                        ),
+                    },
+                    {"role": "user", "content": "Now regenerate your response using the reflection you made"},
+                ]
+                new_response = await self._openai_client.async_client.chat.completions.create(
+                    model=self.llm,
+                    messages=regenerate_messages,
+                    temperature=temperature,
+                )
+                response_text = new_response.choices[0].message.content
+                reflection_count += 1
+            except Exception as e:
+                logging.error("Reflection parsing failed.", exc_info=True)
+                reflection_count += 1
+                if reflection_count >= self.max_reflect:
+                    break
+
+        try:
+            validated_response = self._apply_guardrail_with_retry(
+                response_text, original_prompt, temperature, tools,
+                task_name, task_description, task_id,
+            )
+            self._execute_callback_and_display(
+                original_prompt, validated_response, time.time() - start_time,
+                task_name, task_description, task_id,
+            )
+            return await self._atrigger_after_agent_hook(
+                original_prompt, validated_response, start_time,
+            )
+        except Exception as e:
+            logging.error(f"Agent {self.name}: Guardrail validation failed for OpenAI client: {e}")
+            self._truncate_chat_history(chat_history_length)
+            return None
+
     def _execute_callback_and_display(self, prompt: str, response: str, generation_time: float, task_name=None, task_description=None, task_id=None):
         """Helper method to execute callbacks and display interaction.
         
@@ -1934,7 +2079,7 @@ Output MUST be JSON with 'reflection' and 'satisfactory'.
                     
                     # Apply guardrail validation for custom LLM response
                     try:
-                        validated_response = self._apply_guardrail_with_retry(response_text, prompt, temperature, tools, task_name, task_description, task_id)
+                        validated_response = await self._aapply_guardrail_with_retry(response_text, prompt, temperature, tools, task_name, task_description, task_id)
                         # Execute callback after validation
                         self._execute_callback_and_display(normalized_content, validated_response, time.time() - start_time, task_name, task_description, task_id)
                         return await self._atrigger_after_agent_hook(prompt, validated_response, start_time)
@@ -2042,8 +2187,216 @@ Output MUST be JSON with 'reflection' and 'satisfactory'.
                             task_id=task_id,
                             response_format=response_format
                         )
-                        # Continue to handle structured outputs and other processing
-                        # Don't return immediately - fall through to existing logic
+                        
+                        # Process response - mirror sync _chat_impl behavior (lines ~1544-1602)
+                        if not response:
+                            # Rollback chat history on response failure
+                            self._truncate_chat_history(chat_history_length)
+                            return None
+
+                        # Extract response content using the same method as sync
+                        response_text = self._extract_llm_response_content(response)
+                        if isinstance(response_text, str):
+                            response_text = response_text.strip()
+
+                        # Handle output_json or output_pydantic if specified
+                        if output_json or output_pydantic:
+                            # Add to chat history and return raw response
+                            # User message already added before LLM call via _build_messages
+                            self._append_to_chat_history({"role": "assistant", "content": response_text})
+                            # Persist assistant message to DB
+                            self._persist_message("assistant", response_text)
+                            # Apply guardrail validation even for JSON output
+                            try:
+                                validated_response = await self._aapply_guardrail_with_retry(response_text, original_prompt, temperature, tools, task_name, task_description, task_id)
+                                # Execute callback after validation
+                                self._execute_callback_and_display(original_prompt, validated_response, time.time() - start_time, task_name, task_description, task_id)
+                                return await self._atrigger_after_agent_hook(original_prompt, validated_response, start_time)
+                            except Exception as e:
+                                logging.error(f"Agent {self.name}: Guardrail validation failed for JSON output: {e}")
+                                # Rollback chat history on guardrail failure
+                                self._truncate_chat_history(chat_history_length)
+                                return None
+
+                        # For regular responses (no self-reflection)
+                        if not self.self_reflect:
+                            # User message already added before LLM call via _build_messages
+                            self._append_to_chat_history({"role": "assistant", "content": response_text})
+                            # Persist assistant message to DB (non-reflect path)
+                            self._persist_message("assistant", response_text)
+                            if self.verbose:
+                                logging.debug(f"Agent {self.name} final response: {response_text}")
+                            # Return only reasoning content if reasoning_steps is True
+                            if reasoning_steps and hasattr(response, 'choices') and response.choices and hasattr(response.choices[0].message, 'reasoning_content') and response.choices[0].message.reasoning_content:
+                                # Apply guardrail to reasoning content
+                                try:
+                                    validated_reasoning = await self._aapply_guardrail_with_retry(response.choices[0].message.reasoning_content, original_prompt, temperature, tools, task_name, task_description, task_id)
+                                    # Execute callback after validation
+                                    self._execute_callback_and_display(original_prompt, validated_reasoning, time.time() - start_time, task_name, task_description, task_id)
+                                    return await self._atrigger_after_agent_hook(original_prompt, validated_reasoning, start_time)
+                                except Exception as e:
+                                    logging.error(f"Agent {self.name}: Guardrail validation failed for reasoning content: {e}")
+                                    # Rollback chat history on guardrail failure
+                                    self._truncate_chat_history(chat_history_length)
+                                    return None
+                            else:
+                                # Apply guardrail to regular response content
+                                try:
+                                    validated_response = await self._aapply_guardrail_with_retry(response_text, original_prompt, temperature, tools, task_name, task_description, task_id)
+                                    # Execute callback after validation
+                                    self._execute_callback_and_display(original_prompt, validated_response, time.time() - start_time, task_name, task_description, task_id)
+                                    return await self._atrigger_after_agent_hook(original_prompt, validated_response, start_time)
+                                except Exception as e:
+                                    logging.error(f"Agent {self.name}: Guardrail validation failed: {e}")
+                                    # Rollback chat history on guardrail failure
+                                    self._truncate_chat_history(chat_history_length)
+                                    return None
+                        
+                        # If self-reflection is enabled, implement reflection logic
+                        if self.self_reflect:
+                            # Implement async self-reflection similar to sync path
+                            reflection_prompt = f"""
+Reflect on your previous response: '{response_text}'.
+{self.reflect_prompt if self.reflect_prompt else "Identify any flaws, improvements, or actions."}
+Provide a "satisfactory" status ('yes' or 'no').
+Output MUST be JSON with 'reflection' and 'satisfactory'.
+                            """
+                            
+                            # Add reflection prompt to messages
+                            reflection_messages = messages + [
+                                {"role": "assistant", "content": response_text},
+                                {"role": "user", "content": reflection_prompt}
+                            ]
+                            
+                            reflection_count = 0
+                            
+                            while True:
+                                try:
+                                    # Check if OpenAI client is available for self-reflection
+                                    if self._openai_client is None:
+                                        # For custom LLMs, self-reflection with structured output is not supported
+                                        if self.verbose:
+                                            _get_display_functions()['display_self_reflection'](f"Agent {self.name}: Self-reflection with structured output is not supported for custom LLM providers. Skipping reflection.", console=self.console)
+                                        # Return the original response without reflection
+                                        self._append_to_chat_history({"role": "assistant", "content": response_text})
+                                        # Persist assistant message to DB
+                                        self._persist_message("assistant", response_text)
+                                        try:
+                                            validated_response = await self._aapply_guardrail_with_retry(response_text, original_prompt, temperature, tools, task_name, task_description, task_id)
+                                            # Execute callback after validation
+                                            self._execute_callback_and_display(original_prompt, validated_response, time.time() - start_time, task_name, task_description, task_id)
+                                            return await self._atrigger_after_agent_hook(original_prompt, validated_response, start_time)
+                                        except Exception as e:
+                                            logging.error(f"Agent {self.name}: Guardrail validation failed: {e}")
+                                            # Rollback chat history on guardrail failure
+                                            self._truncate_chat_history(chat_history_length)
+                                            return None
+                                    
+                                    reflection_response = await self._openai_client.async_client.beta.chat.completions.parse(
+                                        model=self.reflect_llm if self.reflect_llm else self.llm,
+                                        messages=reflection_messages,
+                                        temperature=temperature,
+                                        response_format=_get_display_functions()['ReflectionOutput']
+                                    )
+                                    
+                                    reflection_output = reflection_response.choices[0].message.parsed
+                                    
+                                    if self.verbose:
+                                        _get_display_functions()['display_self_reflection'](f"Agent {self.name} self reflection (using {self.reflect_llm if self.reflect_llm else self.llm}): reflection='{reflection_output.reflection}' satisfactory='{reflection_output.satisfactory}'", console=self.console)
+                                    
+                                    # Only consider satisfactory after minimum reflections
+                                    if reflection_output.satisfactory == "yes" and reflection_count >= self.min_reflect - 1:
+                                        if self.verbose:
+                                            _get_display_functions()['display_self_reflection']("Agent marked the response as satisfactory after meeting minimum reflections", console=self.console)
+                                        # Add to chat history and return
+                                        self._append_to_chat_history({"role": "assistant", "content": response_text})
+                                        # Persist assistant message to DB
+                                        self._persist_message("assistant", response_text)
+                                        try:
+                                            validated_response = await self._aapply_guardrail_with_retry(response_text, original_prompt, temperature, tools, task_name, task_description, task_id)
+                                            # Execute callback after validation
+                                            self._execute_callback_and_display(original_prompt, validated_response, time.time() - start_time, task_name, task_description, task_id)
+                                            return await self._atrigger_after_agent_hook(original_prompt, validated_response, start_time)
+                                        except Exception as e:
+                                            logging.error(f"Agent {self.name}: Guardrail validation failed after reflection: {e}")
+                                            # Rollback chat history on guardrail failure
+                                            self._truncate_chat_history(chat_history_length)
+                                            return None
+                                    
+                                    # Check if we've hit max reflections
+                                    if reflection_count >= self.max_reflect - 1:
+                                        if self.verbose:
+                                            _get_display_functions()['display_self_reflection']("Maximum reflection count reached, returning current response", console=self.console)
+                                        # Add to chat history and return
+                                        self._append_to_chat_history({"role": "assistant", "content": response_text})
+                                        # Persist assistant message to DB
+                                        self._persist_message("assistant", response_text)
+                                        try:
+                                            validated_response = await self._aapply_guardrail_with_retry(response_text, original_prompt, temperature, tools, task_name, task_description, task_id)
+                                            # Execute callback after validation
+                                            self._execute_callback_and_display(original_prompt, validated_response, time.time() - start_time, task_name, task_description, task_id)
+                                            return await self._atrigger_after_agent_hook(original_prompt, validated_response, start_time)
+                                        except Exception as e:
+                                            logging.error(f"Agent {self.name}: Guardrail validation failed after max reflections: {e}")
+                                            # Rollback chat history on guardrail failure
+                                            self._truncate_chat_history(chat_history_length)
+                                            return None
+                                    
+                                    # Regenerate response based on reflection
+                                    regenerate_messages = reflection_messages + [
+                                        {"role": "assistant", "content": f"Self Reflection: {reflection_output.reflection} Satisfactory?: {reflection_output.satisfactory}"},
+                                        {"role": "user", "content": "Now regenerate your response using the reflection you made"}
+                                    ]
+                                    
+                                    new_response = await self._execute_unified_achat_completion(
+                                        messages=regenerate_messages,
+                                        temperature=temperature,
+                                        tools=formatted_tools,
+                                        stream=stream,
+                                        reasoning_steps=reasoning_steps,
+                                        task_name=task_name,
+                                        task_description=task_description,
+                                        task_id=task_id,
+                                        response_format=response_format
+                                    )
+                                    
+                                    if new_response:
+                                        new_response_text = self._extract_llm_response_content(new_response)
+                                        if isinstance(new_response_text, str):
+                                            response_text = new_response_text.strip()
+                                        # Update reflection_messages to include the new response for next iteration
+                                        reflection_messages = regenerate_messages + [
+                                            {"role": "assistant", "content": response_text}
+                                        ]
+                                    
+                                    reflection_count += 1
+                                    
+                                except Exception as e:
+                                    if self.verbose:
+                                        _get_display_functions()['display_error'](f"Error in parsing self-reflection json {e}. Retrying", console=self.console)
+                                    logging.error("Reflection parsing failed.", exc_info=True)
+                                    reflection_count += 1
+                                    if reflection_count >= self.max_reflect:
+                                        # Return original response after max reflection attempts
+                                        self._append_to_chat_history({"role": "assistant", "content": response_text})
+                                        # Persist assistant message to DB
+                                        self._persist_message("assistant", response_text)
+                                        try:
+                                            validated_response = await self._aapply_guardrail_with_retry(response_text, original_prompt, temperature, tools, task_name, task_description, task_id)
+                                            # Execute callback after validation
+                                            self._execute_callback_and_display(original_prompt, validated_response, time.time() - start_time, task_name, task_description, task_id)
+                                            return await self._atrigger_after_agent_hook(original_prompt, validated_response, start_time)
+                                        except Exception as guard_e:
+                                            logging.error(f"Agent {self.name}: Guardrail validation failed after reflection error: {guard_e}")
+                                            # Rollback chat history on guardrail failure
+                                            self._truncate_chat_history(chat_history_length)
+                                            return None
+                                    continue
+                        
+                        # This should never be reached due to the returns above
+                        # But adding as safety fallback
+                        logging.warning(f"Agent {self.name}: Unexpected code path reached in unified dispatch")
+                        return None
                     else:
                         # LEGACY: Check if OpenAI client is available
                         if self._openai_client is None:
@@ -2180,7 +2533,7 @@ Output MUST be JSON with 'reflection' and 'satisfactory'.
                         
                         # Apply guardrail validation for OpenAI client response
                         try:
-                            validated_response = self._apply_guardrail_with_retry(response_text, original_prompt, temperature, tools, task_name, task_description, task_id)
+                            validated_response = await self._aapply_guardrail_with_retry(response_text, original_prompt, temperature, tools, task_name, task_description, task_id)
                             # Execute callback after validation
                             self._execute_callback_and_display(original_prompt, validated_response, time.time() - start_time, task_name, task_description, task_id)
                             return await self._atrigger_after_agent_hook(original_prompt, validated_response, start_time)

src/praisonai-agents/praisonaiagents/agent/code_agent.py

@@ -164,7 +164,8 @@ Follow best practices and coding standards."""
             messages=messages,
             **{k: v for k, v in kwargs.items() if k != "model"}
         )
-        
+        if not response.choices or response.choices[0].message is None:
+            raise ValueError("LLM returned empty or filtered response")
         code = response.choices[0].message.content
         self._log("Code generated successfully", "green")
         return code
@@ -204,7 +205,8 @@ Follow best practices and coding standards."""
             messages=messages,
             **{k: v for k, v in kwargs.items() if k != "model"}
         )
-        
+        if not response.choices or response.choices[0].message is None:
+            raise ValueError("LLM returned empty or filtered response")
         code = response.choices[0].message.content
         self._log("Code generated successfully", "green")
         return code
@@ -365,6 +367,8 @@ Provide clear, actionable feedback."""
             **{k: v for k, v in kwargs.items() if k != "model"}
         )
         
+        if not response.choices or response.choices[0].message is None:
+            raise ValueError("LLM returned empty or filtered response")
         review = response.choices[0].message.content
         self._log("Review completed", "green")
         return review
@@ -405,6 +409,8 @@ Include:
             **{k: v for k, v in kwargs.items() if k != "model"}
         )
         
+        if not response.choices or response.choices[0].message is None:
+            raise ValueError("LLM returned empty or filtered response")
         explanation = response.choices[0].message.content
         self._log("Explanation completed", "green")
         return explanation
@@ -451,6 +457,8 @@ Only output the refactored code."""
             **{k: v for k, v in kwargs.items() if k != "model"}
         )
         
+        if not response.choices or response.choices[0].message is None:
+            raise ValueError("LLM returned empty or filtered response")
         refactored = response.choices[0].message.content
         self._log("Refactoring completed", "green")
         return refactored
@@ -495,6 +503,8 @@ Ensure the fix:
             **{k: v for k, v in kwargs.items() if k != "model"}
         )
         
+        if not response.choices or response.choices[0].message is None:
+            raise ValueError("LLM returned empty or filtered response")
         fixed = response.choices[0].message.content
         self._log("Fix completed", "green")
         return fixed

src/praisonai-agents/praisonaiagents/agent/sandbox_mixin.py

@@ -0,0 +1,231 @@
+"""
+Sandbox execution mixin for Agent.
+
+Provides safe code execution capabilities to agents via the sandbox framework.
+"""
+
+import logging
+from typing import Optional, Dict, Any, Union
+
+from ..sandbox import (
+    SandboxConfig, 
+    SandboxManager, 
+    SandboxResult,
+    check_code_safety,
+    format_warnings,
+)
+
+logger = logging.getLogger(__name__)
+
+
+class SandboxMixin:
+    """Mixin that adds sandbox execution capabilities to Agent.
+    
+    Provides safe code execution in isolated environments with security pre-checks.
+    """
+    
+    def __init__(self, *args, **kwargs):
+        # Extract sandbox config from kwargs
+        self.sandbox_config: Optional[SandboxConfig] = kwargs.pop('sandbox', None)
+        
+        # Convert boolean True to default config
+        if self.sandbox_config is True:
+            self.sandbox_config = SandboxConfig.subprocess()
+        
+        super().__init__(*args, **kwargs)
+        
+        self._sandbox_manager: Optional[SandboxManager] = None
+    
+    @property
+    def has_sandbox(self) -> bool:
+        """Check if sandbox is configured for this agent."""
+        return self.sandbox_config is not None
+    
+    def get_sandbox_manager(self) -> Optional[SandboxManager]:
+        """Get or create sandbox manager."""
+        if not self.has_sandbox:
+            return None
+        
+        if self._sandbox_manager is None:
+            self._sandbox_manager = SandboxManager(self.sandbox_config)
+        
+        return self._sandbox_manager
+    
+    async def execute_code(
+        self,
+        code: str,
+        language: str = "python",
+        check_security: bool = True,
+        **kwargs
+    ) -> SandboxResult:
+        """Execute code safely in configured sandbox.
+        
+        Args:
+            code: Code to execute
+            language: Programming language (python, bash, etc.)
+            check_security: Whether to run security pre-checks
+            **kwargs: Additional arguments passed to sandbox.execute()
+            
+        Returns:
+            SandboxResult with execution details
+            
+        Raises:
+            RuntimeError: If no sandbox is configured
+        """
+        if not self.has_sandbox:
+            raise RuntimeError(
+                "No sandbox configured. Set sandbox=True or provide SandboxConfig to Agent()"
+            )
+        
+        # Security pre-check if enabled
+        if check_security:
+            warnings = check_code_safety(code, language)
+            if warnings:
+                warning_text = format_warnings(warnings)
+                if self.verbose:
+                    logger.warning(f"Security warnings for code execution:\n{warning_text}")
+                
+                # Store warnings in result metadata
+                kwargs.setdefault('metadata', {})['security_warnings'] = warnings
+        
+        manager = self.get_sandbox_manager()
+        return await manager.run_code(code, language=language, **kwargs)
+    
+    def execute_code_sync(
+        self,
+        code: str,
+        language: str = "python", 
+        check_security: bool = True,
+        **kwargs
+    ) -> SandboxResult:
+        """Synchronous wrapper for execute_code.
+        
+        Args:
+            code: Code to execute
+            language: Programming language
+            check_security: Whether to run security pre-checks
+            **kwargs: Additional arguments
+            
+        Returns:
+            SandboxResult with execution details
+        """
+        from ..approval.utils import run_coroutine_safely
+        return run_coroutine_safely(self.execute_code(code, language, check_security, **kwargs))
+    
+    async def run_shell_command(
+        self,
+        command: Union[str, list],
+        check_security: bool = True,
+        **kwargs
+    ) -> SandboxResult:
+        """Run a shell command in the sandbox.
+        
+        Args:
+            command: Shell command to run
+            check_security: Whether to run security pre-checks
+            **kwargs: Additional arguments
+            
+        Returns:
+            SandboxResult with execution details
+        """
+        if not self.has_sandbox:
+            raise RuntimeError("No sandbox configured")
+        
+        # Convert command to string for security checking
+        if isinstance(command, list):
+            command_str = " ".join(command)
+        else:
+            command_str = command
+        
+        # Security pre-check
+        if check_security:
+            warnings = check_code_safety(command_str, "bash")
+            if warnings:
+                warning_text = format_warnings(warnings)
+                if self.verbose:
+                    logger.warning(f"Security warnings for command:\n{warning_text}")
+        
+        async with self.get_sandbox_manager() as sandbox:
+            return await sandbox.run_command(command, **kwargs)
+    
+    def get_sandbox_status(self) -> Dict[str, Any]:
+        """Get sandbox status information.
+        
+        Returns:
+            Dictionary with sandbox status details
+        """
+        if not self.has_sandbox:
+            return {"configured": False}
+        
+        status = {
+            "configured": True,
+            "config": self.sandbox_config.to_dict() if self.sandbox_config else None,
+        }
+        
+        manager = self.get_sandbox_manager()
+        if manager:
+            # Get available sandbox types
+            available_types = manager.get_available_types()
+            status["available_types"] = available_types
+            status["current_type"] = self.sandbox_config.sandbox_type
+        
+        return status
+    
+    async def sandbox_cleanup(self):
+        """Clean up sandbox resources."""
+        if self._sandbox_manager:
+            # The context manager handles cleanup, but we can force it here
+            try:
+                await self._sandbox_manager.__aexit__(None, None, None)
+            except Exception as e:
+                logger.warning(f"Error during sandbox cleanup: {e}")
+            finally:
+                self._sandbox_manager = None
+    
+    def _get_code_execution_tools(self):
+        """Get sandbox-based code execution tools for this agent.
+        
+        Returns tools that use the configured sandbox for safe execution.
+        
+        Returns:
+            List of tool functions
+        """
+        if not self.has_sandbox:
+            return []
+        
+        from ..tools import tool
+        
+        @tool
+        def execute_python_code(code: str) -> str:
+            """Execute Python code safely in sandbox.
+            
+            Args:
+                code: Python code to execute
+                
+            Returns:
+                Execution output
+            """
+            result = self.execute_code_sync(code, language="python")
+            if result.success:
+                return result.stdout or result.output
+            else:
+                return f"Error: {result.error or result.stderr}"
+        
+        @tool  
+        def execute_shell_command(command: str) -> str:
+            """Execute shell command safely in sandbox.
+            
+            Args:
+                command: Shell command to execute
+                
+            Returns:
+                Command output
+            """
+            from ..approval.utils import run_coroutine_safely
+            result = run_coroutine_safely(self.run_shell_command(command))
+            if result.success:
+                return result.stdout or result.output
+            else:
+                return f"Error: {result.error or result.stderr}"
+        
+        return [execute_python_code, execute_shell_command]

src/praisonai-agents/praisonaiagents/agents/agents.py

@@ -126,7 +126,24 @@ class _AgentServerRegistry:
         
         thread = threading.Thread(target=run_server, daemon=True)
         thread.start()
-        ready_event.wait(timeout=5.0)  # Deterministic wait instead of sleep(0.5)
+        
+        # Check for configurable timeout via environment variable
+        try:
+            timeout = float(os.environ.get("PRAISONAI_SERVER_READY_TIMEOUT", "5.0"))
+        except ValueError:
+            logger.warning("Invalid PRAISONAI_SERVER_READY_TIMEOUT value. Using default 5.0s.")
+            timeout = 5.0
+        became_ready = ready_event.wait(timeout=timeout)
+        
+        if not became_ready:
+            logger.warning(
+                "Agent server on port %s did not become ready within %.1fs. "
+                "Proceeding, but some features may not work correctly. "
+                "Check server logs for startup errors.",
+                port,
+                timeout,
+            )
+        
         return True
 
 

src/praisonai-agents/praisonaiagents/allowed_tools_filter.py

@@ -0,0 +1,257 @@
+"""
+Tool whitelist filter module with diagnostics.
+
+This module provides a canonical implementation of the ALLOWED_TOOLS filter
+that prevents tool name collisions in multi-environment agent systems by 
+whitelisting only specified tools.
+
+The filter is designed to solve the "tool shadowing" problem where multiple
+modules register tools with overlapping or ambiguous names, causing agents
+to invoke the wrong implementation.
+
+Environment Variables:
+    ALLOWED_TOOLS: Primary variable (comma-separated tool names)
+    HERMES_ONLY_TOOLS: Backward compatibility alias for ALLOWED_TOOLS
+
+Usage:
+    from praisonaiagents.allowed_tools_filter import AllowedToolsFilter
+    
+    # Create filter from environment variable  
+    filter = AllowedToolsFilter()
+    
+    # Filter tools
+    available_tools = {"search", "send_message", "extract_pdf"}
+    filtered_tools = filter.filter_tools(available_tools)
+    
+    # Get diagnostics
+    filter.log_diagnostics()
+"""
+
+import os
+import logging
+from typing import Any, Dict, List, Optional, Set, Union
+
+logger = logging.getLogger(__name__)
+
+
+class AllowedToolsFilter:
+    """
+    Canonical tool whitelist filter with diagnostics and backward compatibility.
+    
+    This filter implements the semantics for tool filtering:
+    - ALLOWED_TOOLS unset (or HERMES_ONLY_TOOLS): all tools visible (with warning about collisions)
+    - ALLOWED_TOOLS empty string: error - must specify tools or unset
+    - ALLOWED_TOOLS with values: only whitelisted tools visible
+    - Unknown tools in list: warn and strip in dev, strict fail in CI
+    
+    Backward Compatibility:
+    - Supports both ALLOWED_TOOLS and HERMES_ONLY_TOOLS
+    - ALLOWED_TOOLS takes precedence if both are set
+    """
+    
+    def __init__(self, env_var_name: str = "ALLOWED_TOOLS"):
+        """
+        Initialize the filter with backward compatibility support.
+        
+        Args:
+            env_var_name: Primary environment variable name (default: "ALLOWED_TOOLS")
+        """
+        # Support both new (ALLOWED_TOOLS) and legacy (HERMES_ONLY_TOOLS) naming
+        self.primary_var = env_var_name
+        self.legacy_var = "HERMES_ONLY_TOOLS"
+        
+        # ALLOWED_TOOLS takes precedence over HERMES_ONLY_TOOLS for backward compatibility
+        self.env_value = os.environ.get(self.primary_var) or os.environ.get(self.legacy_var)
+        self.env_var_name = self.primary_var if os.environ.get(self.primary_var) else self.legacy_var
+        
+        self.is_ci = os.environ.get("CI", "").lower() in ("true", "1", "yes")
+        self._whitelist: Optional[Set[str]] = self._parse_whitelist()
+        self._diagnostics: Dict[str, Any] = {}
+        
+    def _parse_whitelist(self) -> Optional[Set[str]]:
+        """
+        Parse the environment variable value into a whitelist set.
+        
+        Returns:
+            None if unset, Set of tool names if set, raises ValueError for empty string
+        """
+        if self.env_value is None:
+            return None
+        
+        # Empty string is an error - must be explicit
+        if self.env_value.strip() == "":
+            raise ValueError(
+                f"{self.env_var_name} cannot be empty. Either unset it for all tools "
+                "or provide a comma-separated list of tool names."
+            )
+        
+        # Parse comma-separated values, strip whitespace
+        tools = {name.strip() for name in self.env_value.split(",") if name.strip()}
+        
+        if not tools:
+            raise ValueError(
+                f"{self.env_var_name} contains no valid tool names. "
+                "Provide comma-separated tool names or unset the variable."
+            )
+        
+        return tools
+    
+    def filter_tools(self, available_tools: Union[Set[str], List[str], Dict[str, Any]]) -> Set[str]:
+        """
+        Filter tools based on ALLOWED_TOOLS whitelist.
+        
+        Args:
+            available_tools: Available tools as set, list, or dict with tool names as keys
+            
+        Returns:
+            Filtered set of tool names
+            
+        Raises:
+            ValueError: If unknown tools in CI mode or configuration errors
+        """
+        # Convert input to set of tool names
+        if isinstance(available_tools, dict):
+            tool_names = set(available_tools.keys())
+        else:
+            tool_names = set(available_tools)
+        
+        # Store for diagnostics
+        self._diagnostics["registered_before_filter"] = sorted(tool_names)
+        
+        # If no whitelist, return all tools (with warning)
+        if self._whitelist is None:
+            logger.warning(
+                "Tool whitelist is unset. All %d tools are visible. "
+                "Consider using ALLOWED_TOOLS to prevent tool name collisions.",
+                len(tool_names)
+            )
+            self._diagnostics["registered_after_filter"] = sorted(tool_names)
+            self._diagnostics["dropped_tools"] = []
+            self._diagnostics["unknown_tools"] = []
+            return tool_names
+        
+        # Find intersection and unknown tools
+        available_whitelisted = tool_names.intersection(self._whitelist)
+        unknown_tools = self._whitelist.difference(tool_names)
+        dropped_tools = tool_names.difference(self._whitelist)
+        
+        # Store for diagnostics
+        self._diagnostics["registered_after_filter"] = sorted(available_whitelisted)
+        self._diagnostics["dropped_tools"] = sorted(dropped_tools)
+        self._diagnostics["unknown_tools"] = sorted(unknown_tools)
+        
+        # Handle unknown tools based on environment
+        if unknown_tools:
+            unknown_list = sorted(unknown_tools)
+            if self.is_ci:
+                raise ValueError(
+                    f"Unknown tools in {self.env_var_name}: {unknown_list}. "
+                    "All specified tools must be available in CI mode."
+                )
+            else:
+                logger.warning(
+                    "Unknown tools in %s will be ignored: %s",
+                    self.env_var_name,
+                    unknown_list
+                )
+        
+        if dropped_tools:
+            logger.info(
+                "Dropped %d tools not in %s whitelist: %s",
+                len(dropped_tools),
+                self.env_var_name,
+                sorted(dropped_tools)
+            )
+        
+        logger.info(
+            "%s filtered tools: %d available, %d after filter",
+            self.env_var_name,
+            len(tool_names),
+            len(available_whitelisted)
+        )
+        
+        return available_whitelisted
+    
+    def log_diagnostics(self) -> None:
+        """Log startup diagnostics section as specified in the issue."""
+        if not self._diagnostics:
+            logger.warning("No diagnostics available. Call filter_tools() first.")
+            return
+        
+        logger.info("=" * 50)
+        logger.info("ALLOWED_TOOLS FILTER DIAGNOSTICS")
+        logger.info("=" * 50)
+        logger.info("%s=%s", self.env_var_name, self.env_value or "<unset>")
+        logger.info("RegisteredBeforeFilter=%s", self._diagnostics.get("registered_before_filter", []))
+        logger.info("RegisteredAfterFilter=%s", self._diagnostics.get("registered_after_filter", []))
+        logger.info("DroppedTools=%s", self._diagnostics.get("dropped_tools", []))
+        logger.info("UnknownTools=%s", self._diagnostics.get("unknown_tools", []))
+        logger.info("IsCI=%s", self.is_ci)
+        logger.info("=" * 50)
+    
+    def get_diagnostics(self) -> Dict[str, Any]:
+        """
+        Get diagnostics data as a dictionary.
+        
+        Returns:
+            Dictionary with diagnostics information
+        """
+        return {
+            "env_var_name": self.env_var_name,
+            "env_value": self.env_value,
+            "is_ci": self.is_ci,
+            "whitelist": list(self._whitelist) if self._whitelist else None,
+            **self._diagnostics
+        }
+    
+    def is_enabled(self) -> bool:
+        """
+        Check if filtering is enabled.
+        
+        Returns:
+            True if tool whitelist is set and filtering is active
+        """
+        return self._whitelist is not None
+    
+    def get_whitelist(self) -> Optional[Set[str]]:
+        """
+        Get the current whitelist.
+        
+        Returns:
+            Set of whitelisted tool names, or None if not set
+        """
+        return self._whitelist.copy() if self._whitelist else None
+
+
+def filter_tools_with_allowed_tools(
+    available_tools: Union[Set[str], List[str], Dict[str, Any]],
+    env_var_name: str = "ALLOWED_TOOLS",
+    log_diagnostics: bool = True
+) -> Set[str]:
+    """
+    Convenience function to filter tools with ALLOWED_TOOLS.
+    
+    Args:
+        available_tools: Available tools to filter
+        env_var_name: Environment variable name (default: "ALLOWED_TOOLS")
+        log_diagnostics: Whether to log diagnostics (default: True)
+        
+    Returns:
+        Filtered set of tool names
+    """
+    filter_instance = AllowedToolsFilter(env_var_name)
+    filtered = filter_instance.filter_tools(available_tools)
+    
+    if log_diagnostics:
+        filter_instance.log_diagnostics()
+    
+    return filtered
+
+
+# Backward compatibility aliases
+filter_tools_with_hermes = filter_tools_with_allowed_tools
+hermes_filter = filter_tools_with_allowed_tools
+apply_hermes_filter = filter_tools_with_allowed_tools
+
+# Legacy class alias for backward compatibility
+HermesToolFilter = AllowedToolsFilter

src/praisonai-agents/praisonaiagents/context/session_tracker.py

@@ -7,7 +7,7 @@ Inspired by Agno's SessionContextStore pattern.
 
 from dataclasses import dataclass, field, asdict
 from typing import Any, Dict, List, Optional
-from datetime import datetime
+from datetime import datetime, timezone
 import json
 import logging
 from praisonaiagents._logging import get_logger
@@ -97,8 +97,8 @@ class SessionContextTracker:
         
         self._state = SessionState(
             session_id=self.session_id,
-            created_at=datetime.utcnow().isoformat() + "Z",
-            updated_at=datetime.utcnow().isoformat() + "Z",
+            created_at=datetime.now(timezone.utc).isoformat(),
+            updated_at=datetime.now(timezone.utc).isoformat(),
         )
     
     def _generate_session_id(self) -> str:
@@ -163,7 +163,7 @@ class SessionContextTracker:
     
     def _mark_updated(self) -> None:
         """Update the updated_at timestamp."""
-        self._state.updated_at = datetime.utcnow().isoformat() + "Z"
+        self._state.updated_at = datetime.now(timezone.utc).isoformat()
         self._state.turn_count += 1
     
     def to_context_string(self) -> str:
@@ -257,7 +257,7 @@ Current messages take precedence if there's any conflict with this summary.
         self._state = SessionState(
             session_id=self.session_id,
             created_at=self._state.created_at,
-            updated_at=datetime.utcnow().isoformat() + "Z",
+            updated_at=datetime.now(timezone.utc).isoformat(),
         )
     
     def to_dict(self) -> Dict[str, Any]:

src/praisonai-agents/praisonaiagents/eval/grader.py

@@ -12,7 +12,7 @@ import os
 import logging
 from praisonaiagents._logging import get_logger
 from dataclasses import dataclass, field, asdict
-from datetime import datetime
+from datetime import datetime, timezone
 from typing import Any, Dict, List, Optional
 
 logger = get_logger(__name__)
@@ -39,7 +39,7 @@ class GradeResult:
     input_text: str = ""
     output: str = ""
     expected_output: Optional[str] = None
-    timestamp: str = field(default_factory=lambda: datetime.utcnow().isoformat())
+    timestamp: str = field(default_factory=lambda: datetime.now(timezone.utc).isoformat())
     
     def to_dict(self) -> Dict[str, Any]:
         """Convert to dictionary."""
@@ -55,7 +55,7 @@ class GradeResult:
             input_text=data.get("input_text", ""),
             output=data.get("output", ""),
             expected_output=data.get("expected_output"),
-            timestamp=data.get("timestamp", datetime.utcnow().isoformat()),
+            timestamp=data.get("timestamp", datetime.now(timezone.utc).isoformat()),
         )
 
 class BaseLLMGrader:

src/praisonai-agents/praisonaiagents/llm/openai_client.py

@@ -907,6 +907,8 @@ class OpenAIClient:
                     final_response = self._responses_to_chat_completion(raw)
                     
                     # Display the response text if display_fn provided
+                    if not final_response.choices or final_response.choices[0].message is None:
+                        raise ValueError("LLM returned empty or filtered response")
                     response_text = final_response.choices[0].message.content or ""
                     if display_fn and response_text:
                         from rich.markdown import Markdown
@@ -1167,6 +1169,8 @@ class OpenAIClient:
                     final_response = self._responses_to_chat_completion(raw)
                     
                     # Display the response text if display_fn provided
+                    if not final_response.choices or final_response.choices[0].message is None:
+                        raise ValueError("LLM returned empty or filtered response")
                     response_text = final_response.choices[0].message.content or ""
                     if display_fn and response_text:
                         from rich.markdown import Markdown
@@ -1638,8 +1642,10 @@ class OpenAIClient:
             )
             
             # Check for tool calls
+            if not final_response.choices or final_response.choices[0].message is None:
+                raise ValueError("LLM returned empty or filtered response")
             tool_calls = getattr(final_response.choices[0].message, 'tool_calls', None)
-            
+
             # Emit llm_content for intermediate narrative display
             # (gpt-4.1+ models produce text alongside tool calls)
             response_content = getattr(final_response.choices[0].message, 'content', None)
@@ -1835,10 +1841,12 @@ class OpenAIClient:
             
             if not final_response:
                 return None
-            
+
             # Check for tool calls
+            if not final_response.choices or final_response.choices[0].message is None:
+                raise ValueError("LLM returned empty or filtered response")
             tool_calls = getattr(final_response.choices[0].message, 'tool_calls', None)
-            
+
             # Emit llm_content for intermediate narrative display
             # (gpt-4.1+ models produce text alongside tool calls)
             response_content = getattr(final_response.choices[0].message, 'content', None)
@@ -2037,6 +2045,8 @@ class OpenAIClient:
                     return
                 
                 # Check for tool calls
+                if not final_response.choices or final_response.choices[0].message is None:
+                    raise ValueError("LLM returned empty or filtered response")
                 tool_calls = getattr(final_response.choices[0].message, 'tool_calls', None)
                 
                 if tool_calls and execute_tool_fn:

src/praisonai-agents/praisonaiagents/memory/memory.py

@@ -7,7 +7,7 @@ import threading
 from typing import Any, Dict, List, Optional, Union, Literal
 import logging
 from praisonaiagents._logging import get_logger
-from datetime import datetime
+from datetime import datetime, timezone
 
 # Decomposed memory functionality - imported as mixins for backward compatibility
 from .storage import StorageMixin
@@ -692,7 +692,7 @@ class Memory(StorageMixin, SearchMixin, MemoryCoreMixin):
                         "_id": ident,
                         "content": text,
                         "metadata": metadata,
-                        "created_at": datetime.utcnow(),
+                        "created_at": datetime.now(timezone.utc),
                         "memory_type": "short_term"
                     }
                     self.mongo_short_term.insert_one(doc)
@@ -918,7 +918,7 @@ class Memory(StorageMixin, SearchMixin, MemoryCoreMixin):
                     "_id": ident,
                     "content": text,
                     "metadata": metadata,
-                    "created_at": datetime.utcnow(),
+                    "created_at": datetime.now(timezone.utc),
                     "memory_type": "long_term"
                 }
                 
@@ -1441,14 +1441,14 @@ class Memory(StorageMixin, SearchMixin, MemoryCoreMixin):
             self.mem0_client.add(text, user_id=user_id, metadata=meta)
         elif self.use_mongodb and hasattr(self, "mongo_users"):
             try:
-                from datetime import datetime
+                from datetime import datetime, timezone
                 ident = str(time.time_ns())
                 doc = {
                     "_id": ident,
                     "user_id": user_id,
                     "content": text,
                     "metadata": meta,
-                    "created_at": datetime.utcnow()
+                    "created_at": datetime.now(timezone.utc)
                 }
                 self.mongo_users.insert_one(doc)
                 self._log_verbose(f"Successfully stored user memory for {user_id}")

src/praisonai-agents/praisonaiagents/sandbox/__init__.py

@@ -15,6 +15,8 @@ from .protocols import (
     ResourceLimits,
 )
 from .config import SandboxConfig, SecurityPolicy
+from .manager import SandboxManager
+from .security import SecurityWarning, check_code_safety, format_warnings, get_security_summary
 
 __all__ = [
     "SandboxProtocol",
@@ -22,5 +24,10 @@ __all__ = [
     "SandboxStatus",
     "ResourceLimits",
     "SandboxConfig",
-    "SecurityPolicy",
+    "SecurityPolicy", 
+    "SandboxManager",
+    "SecurityWarning",
+    "check_code_safety",
+    "format_warnings", 
+    "get_security_summary",
 ]

src/praisonai-agents/praisonaiagents/sandbox/manager.py

@@ -0,0 +1,302 @@
+"""
+SandboxManager for PraisonAI Agents.
+
+Factory and context manager for sandbox backends.
+Core SDK component that routes to appropriate sandbox implementations
+in the praisonai wrapper package.
+"""
+
+from __future__ import annotations
+
+import logging
+from typing import Dict, Any, Optional, TYPE_CHECKING
+
+from .config import SandboxConfig
+from .protocols import SandboxProtocol, SandboxResult
+
+if TYPE_CHECKING:
+    from typing import AsyncContextManager
+
+logger = logging.getLogger(__name__)
+
+
+class SandboxManager:
+    """Factory and context manager for all sandbox backends.
+    
+    Routes to appropriate sandbox implementations based on configuration.
+    Lightweight manager in core SDK - heavy implementations in wrapper.
+    
+    Example:
+        from praisonaiagents.sandbox import SandboxManager, SandboxConfig
+        
+        # Simple usage
+        config = SandboxConfig.docker("python:3.11-slim")
+        manager = SandboxManager(config)
+        result = await manager.run_code("print('Hello, World!')")
+        
+        # Context manager usage  
+        async with SandboxManager(config) as sandbox:
+            result = await sandbox.execute("print('Hello, World!')")
+    """
+    
+    def __init__(self, config: Optional[SandboxConfig] = None):
+        """Initialize the sandbox manager.
+        
+        Args:
+            config: Sandbox configuration. Defaults to subprocess sandbox.
+        """
+        self.config = config or SandboxConfig.subprocess()
+        self._sandbox: Optional[SandboxProtocol] = None
+        
+    async def __aenter__(self) -> SandboxProtocol:
+        """Async context manager entry."""
+        self._sandbox = await self._create_sandbox()
+        return self._sandbox
+        
+    async def __aexit__(self, exc_type, exc_val, exc_tb):
+        """Async context manager exit with cleanup."""
+        if self._sandbox:
+            try:
+                await self._sandbox.stop()
+                await self._sandbox.cleanup()
+            except Exception as e:
+                logger.warning(f"Error during sandbox cleanup: {e}")
+            finally:
+                self._sandbox = None
+    
+    async def run_code(
+        self, 
+        code: str,
+        language: str = "python",
+        **kwargs
+    ) -> SandboxResult:
+        """Convenience method: create sandbox, run code, cleanup.
+        
+        Args:
+            code: Code to execute
+            language: Programming language
+            **kwargs: Additional arguments passed to execute()
+            
+        Returns:
+            SandboxResult with execution details
+        """
+        async with self as sandbox:
+            return await sandbox.execute(code, language=language, **kwargs)
+    
+    async def _create_sandbox(self) -> SandboxProtocol:
+        """Create appropriate sandbox backend based on config.
+        
+        Routes to implementations in praisonai wrapper package.
+        
+        Returns:
+            Configured sandbox instance
+            
+        Raises:
+            ValueError: For unknown sandbox types
+            ImportError: If required dependencies not available
+        """
+        sandbox_type = self.config.sandbox_type.lower()
+        
+        if sandbox_type == "docker":
+            return await self._create_docker_sandbox()
+        elif sandbox_type in ("subprocess", "local"):
+            return await self._create_subprocess_sandbox()  
+        elif sandbox_type == "e2b":
+            return await self._create_e2b_sandbox()
+        elif sandbox_type == "sandlock":
+            return await self._create_sandlock_sandbox()
+        elif sandbox_type == "ssh":
+            return await self._create_ssh_sandbox()
+        elif sandbox_type == "modal":
+            return await self._create_modal_sandbox()
+        elif sandbox_type == "daytona":
+            return await self._create_daytona_sandbox()
+        else:
+            raise ValueError(
+                f"Unknown sandbox type: {sandbox_type!r}. "
+                f"Supported: 'docker', 'subprocess', 'e2b', 'sandlock', 'ssh', 'modal', 'daytona'"
+            )
+    
+    async def _create_docker_sandbox(self) -> SandboxProtocol:
+        """Create Docker sandbox."""
+        try:
+            # Lazy import from wrapper package
+            from praisonai.sandbox import DockerSandbox
+        except ImportError as e:
+            raise ImportError(
+                "Docker sandbox not available. Install with: "
+                "pip install praisonaiagents[sandbox-docker]"
+            ) from e
+        
+        sandbox = DockerSandbox(
+            image=self.config.image,
+            config=self.config,
+        )
+        
+        if not sandbox.is_available:
+            raise RuntimeError(
+                "Docker is not available. Please install Docker and ensure it's running."
+            )
+        
+        await sandbox.start()
+        return sandbox
+    
+    async def _create_subprocess_sandbox(self) -> SandboxProtocol:
+        """Create subprocess sandbox."""
+        try:
+            from praisonai.sandbox import SubprocessSandbox
+        except ImportError as e:
+            raise ImportError(
+                "Subprocess sandbox not available. This should not happen - "
+                "please check your installation."
+            ) from e
+        
+        sandbox = SubprocessSandbox(config=self.config)
+        await sandbox.start()
+        return sandbox
+    
+    async def _create_e2b_sandbox(self) -> SandboxProtocol:
+        """Create E2B sandbox."""
+        try:
+            from praisonai.sandbox import E2BSandbox
+        except ImportError as e:
+            raise ImportError(
+                "E2B sandbox not available. Install with: "
+                "pip install praisonaiagents[sandbox] e2b-code-interpreter"
+            ) from e
+        
+        sandbox = E2BSandbox(config=self.config)
+        
+        if not sandbox.is_available:
+            raise RuntimeError(
+                "E2B is not available. Please set E2B_API_KEY environment variable."
+            )
+        
+        await sandbox.start()
+        return sandbox
+    
+    async def _create_sandlock_sandbox(self) -> SandboxProtocol:
+        """Create Sandlock sandbox.""" 
+        try:
+            from praisonai.sandbox import SandlockSandbox
+        except ImportError as e:
+            raise ImportError(
+                "Sandlock sandbox not available. Install sandlock."
+            ) from e
+        
+        sandbox = SandlockSandbox(config=self.config)
+        await sandbox.start()
+        return sandbox
+        
+    async def _create_ssh_sandbox(self) -> SandboxProtocol:
+        """Create SSH sandbox."""
+        try:
+            from praisonai.sandbox import SSHSandbox
+        except ImportError as e:
+            raise ImportError(
+                "SSH sandbox not available. Install with: "
+                "pip install paramiko"
+            ) from e
+        
+        sandbox = SSHSandbox(config=self.config)
+        await sandbox.start()
+        return sandbox
+        
+    async def _create_modal_sandbox(self) -> SandboxProtocol:
+        """Create Modal sandbox."""
+        try:
+            from praisonai.sandbox import ModalSandbox
+        except ImportError as e:
+            raise ImportError(
+                "Modal sandbox not available. Install with: "
+                "pip install modal"
+            ) from e
+        
+        sandbox = ModalSandbox(config=self.config)
+        await sandbox.start()
+        return sandbox
+        
+    async def _create_daytona_sandbox(self) -> SandboxProtocol:
+        """Create Daytona sandbox."""
+        try:
+            from praisonai.sandbox import DaytonaSandbox
+        except ImportError as e:
+            raise ImportError(
+                "Daytona sandbox not available. Install daytona client."
+            ) from e
+        
+        sandbox = DaytonaSandbox(config=self.config)
+        await sandbox.start()
+        return sandbox
+
+    def get_available_types(self) -> Dict[str, Dict[str, Any]]:
+        """Get available sandbox types and their status.
+        
+        Returns:
+            Dictionary mapping sandbox types to availability info
+        """
+        types = {}
+        
+        # Check Docker
+        try:
+            from praisonai.sandbox import DockerSandbox
+            docker = DockerSandbox()
+            types["docker"] = {
+                "available": docker.is_available,
+                "description": "Isolated Docker containers",
+                "requires": ["docker"],
+            }
+        except ImportError:
+            types["docker"] = {
+                "available": False,
+                "description": "Isolated Docker containers", 
+                "requires": ["docker", "praisonaiagents[sandbox-docker]"],
+            }
+        
+        # Check subprocess (always available)
+        types["subprocess"] = {
+            "available": True,
+            "description": "Local subprocess (limited isolation)",
+            "requires": [],
+        }
+        
+        # Check E2B
+        try:
+            from praisonai.sandbox import E2BSandbox
+            e2b = E2BSandbox()
+            types["e2b"] = {
+                "available": e2b.is_available,
+                "description": "E2B cloud sandboxes",
+                "requires": ["e2b-code-interpreter", "E2B_API_KEY"],
+            }
+        except ImportError:
+            types["e2b"] = {
+                "available": False,
+                "description": "E2B cloud sandboxes",
+                "requires": ["e2b-code-interpreter", "E2B_API_KEY"],
+            }
+        
+        # Add other types with try/except blocks
+        for sandbox_type, module_name, description, requirements in [
+            ("sandlock", "SandlockSandbox", "OS-native sandboxing", ["sandlock"]),
+            ("ssh", "SSHSandbox", "Remote SSH execution", ["paramiko"]),
+            ("modal", "ModalSandbox", "Modal cloud compute", ["modal"]),
+            ("daytona", "DaytonaSandbox", "Daytona workspaces", ["daytona"]),
+        ]:
+            try:
+                from praisonai import sandbox
+                sandbox_class = getattr(sandbox, module_name)
+                # Check if class exists without instantiating (some need required args)
+                types[sandbox_type] = {
+                    "available": False,  # Can't easily check without required args
+                    "description": description,
+                    "requires": requirements,
+                }
+            except (ImportError, AttributeError):
+                types[sandbox_type] = {
+                    "available": False,
+                    "description": description,
+                    "requires": requirements,
+                }
+        
+        return types

src/praisonai-agents/praisonaiagents/sandbox/security.py

@@ -0,0 +1,345 @@
+"""
+Security pre-checks for sandbox code execution.
+
+Provides pre-execution analysis of code for potentially dangerous patterns.
+Note: These are warnings only - the sandbox provides the real isolation.
+"""
+
+import re
+import ast
+import logging
+from typing import List, Optional, Dict, Any
+from dataclasses import dataclass
+
+logger = logging.getLogger(__name__)
+
+
+@dataclass 
+class SecurityWarning:
+    """A security warning from code analysis."""
+    
+    pattern: str
+    message: str
+    severity: str  # "low", "medium", "high", "critical"
+    line_number: Optional[int] = None
+    context: Optional[str] = None
+
+
+# Dangerous patterns to detect in code
+DANGEROUS_PATTERNS = [
+    # System/OS access
+    (r"import\s+os\s*;?\s*os\.system", "Direct system command execution via os.system", "high"),
+    (r"subprocess\.call.*shell\s*=\s*True", "Shell injection risk with subprocess", "high"),
+    (r"subprocess\.run.*shell\s*=\s*True", "Shell injection risk with subprocess", "high"),
+    (r"subprocess\.Popen.*shell\s*=\s*True", "Shell injection risk with subprocess", "high"),
+    (r"__import__\(['\"]os['\"]\)", "Dynamic import of os module", "medium"),
+    (r"eval\s*\(", "Dynamic code execution with eval()", "critical"),
+    (r"exec\s*\(", "Dynamic code execution with exec()", "critical"),
+    (r"compile\s*\(.*exec", "Code compilation for execution", "high"),
+    
+    # File system access
+    (r"open\s*\(.*['\"]\/etc", "Access to system configuration files", "medium"),
+    (r"open\s*\(.*['\"]\/root", "Access to root directory", "high"),
+    (r"open\s*\(.*['\"]\/home", "Access to user directories", "medium"),
+    (r"shutil\.rmtree", "Recursive directory deletion", "high"),
+    (r"os\.remove\s*\(", "File deletion", "medium"),
+    (r"os\.unlink\s*\(", "File deletion", "medium"),
+    (r"pathlib\.Path.*unlink", "File deletion via pathlib", "medium"),
+    
+    # Network access
+    (r"socket\.socket", "Direct socket creation", "medium"),
+    (r"urllib\.request", "HTTP requests via urllib", "low"),
+    (r"requests\.", "HTTP requests via requests library", "low"),
+    (r"http\.client", "HTTP client usage", "low"),
+    
+    # Process manipulation
+    (r"os\.fork\s*\(", "Process forking", "high"),
+    (r"multiprocessing\.", "Multiprocessing usage", "medium"),
+    (r"threading\.Thread", "Thread creation", "low"),
+    
+    # Potentially dangerous builtins
+    (r"getattr\s*\(.*__", "Dynamic attribute access to dunder methods", "medium"),
+    (r"setattr\s*\(.*__", "Dynamic attribute setting of dunder methods", "medium"),
+    (r"hasattr\s*\(.*__", "Dynamic attribute checking of dunder methods", "low"),
+    (r"globals\s*\(\)", "Access to global namespace", "medium"),
+    (r"locals\s*\(\)", "Access to local namespace", "low"),
+    (r"vars\s*\(", "Variable inspection", "low"),
+    (r"dir\s*\(", "Object inspection", "low"),
+    
+    # Code modification
+    (r"sys\.modules", "Module manipulation", "high"),
+    (r"importlib\.", "Dynamic import manipulation", "medium"),
+    (r"__builtins__", "Access to builtin functions", "high"),
+    
+    # Memory/performance
+    (r"while\s+True:", "Infinite loop (potential DoS)", "medium"),
+    (r"for.*in.*range\s*\(\s*[0-9]{6,}", "Large range iteration", "low"),
+    (r"list\s*\(\s*range\s*\(\s*[0-9]{6,}", "Large list creation", "low"),
+    (r"\*\s*[0-9]{6,}", "Large data multiplication", "low"),
+]
+
+
+def check_code_safety(code: str, language: str = "python") -> List[SecurityWarning]:
+    """Analyze code for potentially dangerous patterns.
+    
+    Args:
+        code: Code to analyze
+        language: Programming language (currently only 'python' supported)
+        
+    Returns:
+        List of security warnings found
+        
+    Note:
+        These are warnings only. The sandbox provides real isolation.
+        This is a best-effort static analysis for user awareness.
+    """
+    warnings = []
+    
+    if language == "python":
+        warnings.extend(_check_python_patterns(code))
+        warnings.extend(_check_python_ast(code))
+    elif language == "bash":
+        warnings.extend(_check_bash_patterns(code))
+    else:
+        # Fall back to basic pattern matching
+        warnings.extend(_check_generic_patterns(code))
+    
+    return warnings
+
+
+def _check_python_patterns(code: str) -> List[SecurityWarning]:
+    """Check Python code using regex patterns."""
+    warnings = []
+    lines = code.split('\n')
+    
+    for i, line in enumerate(lines, 1):
+        for pattern, message, severity in DANGEROUS_PATTERNS:
+            if re.search(pattern, line, re.IGNORECASE):
+                warnings.append(SecurityWarning(
+                    pattern=pattern,
+                    message=message,
+                    severity=severity,
+                    line_number=i,
+                    context=line.strip(),
+                ))
+    
+    return warnings
+
+
+def _check_python_ast(code: str) -> List[SecurityWarning]:
+    """Check Python code using AST analysis."""
+    warnings = []
+    
+    try:
+        tree = ast.parse(code)
+        visitor = DangerousASTVisitor()
+        visitor.visit(tree)
+        warnings.extend(visitor.warnings)
+    except SyntaxError:
+        # Invalid Python syntax - let the sandbox handle it
+        pass
+    except Exception as e:
+        logger.debug(f"AST analysis failed: {e}")
+    
+    return warnings
+
+
+def _check_bash_patterns(code: str) -> List[SecurityWarning]:
+    """Check bash/shell code for dangerous patterns."""
+    warnings = []
+    lines = code.split('\n')
+    
+    bash_patterns = [
+        (r"rm\s+-rf\s+\/", "Recursive deletion of root directory", "critical"),
+        (r"rm\s+-rf\s+\*", "Recursive deletion with wildcard", "high"),
+        (r">\s*\/dev\/sd[a-z]", "Direct disk device access", "critical"),
+        (r"dd\s+if=.*of=", "Low-level disk operations", "high"),
+        (r"mkfs\.", "Filesystem creation", "critical"),
+        (r"fdisk", "Disk partitioning", "critical"),
+        (r"shutdown", "System shutdown", "high"),
+        (r"reboot", "System reboot", "high"),
+        (r"halt", "System halt", "high"),
+        (r"init\s+[0-6]", "System runlevel change", "high"),
+        (r"chmod\s+777", "Overly permissive file permissions", "medium"),
+        (r"curl.*\|\s*bash", "Piped execution from network", "high"),
+        (r"wget.*\|\s*bash", "Piped execution from network", "high"),
+        (r"\$\(.*\)", "Command substitution", "medium"),
+        (r"`.*`", "Command substitution (backticks)", "medium"),
+    ]
+    
+    for i, line in enumerate(lines, 1):
+        for pattern, message, severity in bash_patterns:
+            if re.search(pattern, line, re.IGNORECASE):
+                warnings.append(SecurityWarning(
+                    pattern=pattern,
+                    message=message,
+                    severity=severity,
+                    line_number=i,
+                    context=line.strip(),
+                ))
+    
+    return warnings
+
+
+def _check_generic_patterns(code: str) -> List[SecurityWarning]:
+    """Check code using generic patterns."""
+    warnings = []
+    lines = code.split('\n')
+    
+    generic_patterns = [
+        (r"eval\s*\(", "Dynamic code execution", "critical"),
+        (r"exec\s*\(", "Dynamic code execution", "critical"),
+        (r"system\s*\(", "System command execution", "high"),
+        (r"shell\s*\(", "Shell command execution", "high"),
+    ]
+    
+    for i, line in enumerate(lines, 1):
+        for pattern, message, severity in generic_patterns:
+            if re.search(pattern, line, re.IGNORECASE):
+                warnings.append(SecurityWarning(
+                    pattern=pattern,
+                    message=message,
+                    severity=severity,
+                    line_number=i,
+                    context=line.strip(),
+                ))
+    
+    return warnings
+
+
+class DangerousASTVisitor(ast.NodeVisitor):
+    """AST visitor to detect dangerous Python constructs."""
+    
+    def __init__(self):
+        self.warnings = []
+    
+    def visit_Call(self, node):
+        """Visit function calls."""
+        # Check for dangerous function calls
+        if isinstance(node.func, ast.Name):
+            func_name = node.func.id
+            if func_name in ("eval", "exec", "compile"):
+                self.warnings.append(SecurityWarning(
+                    pattern=f"{func_name}()",
+                    message=f"Dynamic code execution with {func_name}()",
+                    severity="critical",
+                    line_number=node.lineno,
+                ))
+        
+        # Check for attribute calls like os.system
+        elif isinstance(node.func, ast.Attribute):
+            if isinstance(node.func.value, ast.Name):
+                module = node.func.value.id
+                func = node.func.attr
+                
+                if module == "os" and func == "system":
+                    self.warnings.append(SecurityWarning(
+                        pattern="os.system()",
+                        message="Direct system command execution",
+                        severity="high",
+                        line_number=node.lineno,
+                    ))
+                elif module == "subprocess" and func in ("call", "run", "Popen"):
+                    # Check if shell=True is used
+                    for keyword in node.keywords:
+                        if keyword.arg == "shell" and isinstance(keyword.value, ast.Constant):
+                            if keyword.value.value is True:
+                                self.warnings.append(SecurityWarning(
+                                    pattern=f"subprocess.{func}(..., shell=True)",
+                                    message="Shell injection risk with subprocess",
+                                    severity="high",
+                                    line_number=node.lineno,
+                                ))
+        
+        self.generic_visit(node)
+    
+    def visit_Import(self, node):
+        """Visit import statements."""
+        for alias in node.names:
+            if alias.name in ("os", "subprocess", "sys"):
+                self.warnings.append(SecurityWarning(
+                    pattern=f"import {alias.name}",
+                    message=f"Import of potentially dangerous module: {alias.name}",
+                    severity="low",
+                    line_number=node.lineno,
+                ))
+        
+        self.generic_visit(node)
+    
+    def visit_ImportFrom(self, node):
+        """Visit from ... import statements."""
+        if node.module in ("os", "subprocess", "sys"):
+            for alias in node.names:
+                if alias.name in ("system", "call", "run", "Popen", "modules"):
+                    self.warnings.append(SecurityWarning(
+                        pattern=f"from {node.module} import {alias.name}",
+                        message=f"Import of dangerous function: {alias.name}",
+                        severity="medium",
+                        line_number=node.lineno,
+                    ))
+        
+        self.generic_visit(node)
+
+
+def format_warnings(warnings: List[SecurityWarning]) -> str:
+    """Format security warnings for display.
+    
+    Args:
+        warnings: List of security warnings
+        
+    Returns:
+        Formatted string of warnings
+    """
+    if not warnings:
+        return "No security issues detected."
+    
+    # Group by severity
+    by_severity = {"critical": [], "high": [], "medium": [], "low": []}
+    for warning in warnings:
+        by_severity[warning.severity].append(warning)
+    
+    lines = []
+    lines.append(f"Security analysis found {len(warnings)} potential issue(s):")
+    lines.append("")
+    
+    for severity in ["critical", "high", "medium", "low"]:
+        if by_severity[severity]:
+            lines.append(f"{severity.upper()} RISK:")
+            for warning in by_severity[severity]:
+                line_info = f" (line {warning.line_number})" if warning.line_number else ""
+                context_info = f"\n  Context: {warning.context}" if warning.context else ""
+                lines.append(f"  - {warning.message}{line_info}{context_info}")
+            lines.append("")
+    
+    lines.append("Note: These are warnings only. The sandbox provides real isolation.")
+    return "\n".join(lines)
+
+
+def get_security_summary(warnings: List[SecurityWarning]) -> Dict[str, Any]:
+    """Get security analysis summary.
+    
+    Args:
+        warnings: List of security warnings
+        
+    Returns:
+        Dictionary with summary statistics
+    """
+    by_severity = {"critical": 0, "high": 0, "medium": 0, "low": 0}
+    for warning in warnings:
+        by_severity[warning.severity] += 1
+    
+    max_severity = "low"
+    if by_severity["critical"] > 0:
+        max_severity = "critical"
+    elif by_severity["high"] > 0:
+        max_severity = "high"  
+    elif by_severity["medium"] > 0:
+        max_severity = "medium"
+    
+    return {
+        "total_warnings": len(warnings),
+        "by_severity": by_severity,
+        "max_severity": max_severity,
+        "is_safe": len(warnings) == 0,
+        "has_critical": by_severity["critical"] > 0,
+    }

src/praisonai-agents/praisonaiagents/skills/manager.py

@@ -6,7 +6,7 @@ from typing import List, Optional, Dict
 logger = logging.getLogger(__name__)
 
 from .models import SkillMetadata
-from .discovery import discover_skills
+from .discovery import discover_skills, get_default_skill_dirs
 from .loader import SkillLoader, LoadedSkill
 from .prompt import generate_skills_xml
 from .substitution import render_skill_body
@@ -271,8 +271,7 @@ class SkillManager:
                 return {"success": False, "error": "Skill content exceeds maximum size (100KB)"}
             
             # Create skill directory
-            from .discovery import get_default_skill_directories
-            skill_dirs = get_default_skill_directories()
+            skill_dirs = get_default_skill_dirs()
             base_dir = skill_dirs[0] if skill_dirs else "~/.praisonai/skills"
             
             import os

src/praisonai-agents/praisonaiagents/tools/__init__.py

@@ -10,7 +10,7 @@ _tools_lazy_cache = {}
 # Export core tool items for organized imports (lightweight)
 from .base import BaseTool, ToolResult, ToolValidationError, validate_tool
 from .decorator import tool, FunctionTool
-from .registry import get_registry, register_tool, get_tool, add_tool, has_tool, remove_tool, list_tools, list_available_tools, ToolRegistry
+from .registry import get_registry, register_tool, get_tool, add_tool, has_tool, remove_tool, list_tools, list_available_tools, list_tools_with_allowed_filter, list_tools_with_hermes_filter, ToolRegistry
 from .tools import Tools
 
 # Export Injected type directly for easy access
@@ -75,12 +75,12 @@ TOOL_MAPPINGS = {
 
 
     # Python Tools
-    'execute_code': ('.python_tools', 'PythonTools'),
-    'analyze_code': ('.python_tools', 'PythonTools'),
-    'format_code': ('.python_tools', 'PythonTools'),
-    'lint_code': ('.python_tools', 'PythonTools'),
-    'disassemble_code': ('.python_tools', 'PythonTools'),
-    'python_tools': ('.python_tools', 'PythonTools'),
+    'execute_code': ('.python_tools', None),
+    'analyze_code': ('.python_tools', None),
+    'format_code': ('.python_tools', None),
+    'lint_code': ('.python_tools', None),
+    'disassemble_code': ('.python_tools', None),
+    'python_tools': ('.python_tools', None),
 
 
     # Chain of Thought Training Tools
@@ -326,7 +326,7 @@ def __getattr__(name: str) -> Any:
         if name in [
             'duckduckgo', 'internet_search', 'searxng_search', 'searxng',
             'scrape_page', 'extract_links', 'crawl', 'extract_text',
-            'execute_command', 'list_processes', 'kill_process', 'get_system_info',
+            'execute_command', 'execute_code', 'analyze_code', 'format_code', 'lint_code', 'disassemble_code', 'list_processes', 'kill_process', 'get_system_info',
             'tavily', 'tavily_search', 'tavily_extract', 'tavily_crawl', 'tavily_map',
             'tavily_search_async', 'tavily_extract_async',
             'ydc', 'ydc_search', 'ydc_contents', 'ydc_news', 'ydc_images',
@@ -363,7 +363,7 @@ __all__ = list(TOOL_MAPPINGS.keys()) + [
     'BaseTool', 'ToolResult', 'ToolValidationError', 'validate_tool',
     'tool', 'FunctionTool',
     'get_registry', 'register_tool', 'get_tool', 'add_tool', 'has_tool', 'remove_tool', 
-    'list_tools', 'list_available_tools', 'ToolRegistry',
+    'list_tools', 'list_available_tools', 'list_tools_with_allowed_filter', 'list_tools_with_hermes_filter', 'ToolRegistry',
     'Tools',
     # Validation and retry protocols
     'ValidationResult', 'ToolValidatorProtocol', 'AsyncToolValidatorProtocol', 'PassthroughValidator',

src/praisonai-agents/praisonaiagents/tools/a2ui_tools.py

@@ -0,0 +1,52 @@
+"""
+A2UI tools — optional declarative UI via Google's a2ui-agent-sdk.
+
+Usage:
+    from praisonaiagents import Agent
+    from praisonaiagents.tools.a2ui_tools import send_a2ui_messages
+
+    agent = Agent(name="assistant", tools=[send_a2ui_messages])
+"""
+
+from __future__ import annotations
+
+import json
+from typing import Any, Dict, List, Union
+
+from praisonaiagents.tools.decorator import tool
+
+
+@tool
+def send_a2ui_messages(messages: Union[List[Dict[str, Any]], str]) -> Dict[str, Any]:
+    """
+    Send validated A2UI JSON messages to a connected A2UI renderer client.
+
+    Pass a list of A2UI message dicts (createSurface, updateComponents, etc.)
+    or a JSON string. Requires: pip install praisonaiagents[a2ui]
+
+    Args:
+        messages: A2UI message list or JSON string from the model.
+
+    Returns:
+        Dict with mime_type and serialised messages for A2A/AG-UI clients.
+    """
+    from praisonaiagents.ui.a2ui.adapter import create_a2ui_part
+
+    if isinstance(messages, str):
+        try:
+            messages = json.loads(messages)
+        except json.JSONDecodeError as exc:
+            raise ValueError(f"messages JSON string is not valid JSON: {exc}") from exc
+
+    if not isinstance(messages, list):
+        raise ValueError("messages must be a list of A2UI message dicts or a JSON string")
+
+    part = create_a2ui_part({"messages": messages})
+    data = getattr(getattr(part, "root", part), "data", messages)
+    metadata = getattr(getattr(part, "root", part), "metadata", {}) or {}
+
+    return {
+        "mime_type": metadata.get("mimeType", "application/json+a2ui"),
+        "messages": messages,
+        "a2ui_part": data,
+    }

src/praisonai-agents/praisonaiagents/tools/mentions.py

@@ -268,6 +268,11 @@ class MentionsParser:
     def _process_url_mention(self, url: str) -> Optional[str]:
         """Process @url:https://... mention."""
         try:
+            from praisonaiagents.tools.spider_tools import SpiderTools
+
+            if not SpiderTools()._validate_url(url):
+                return f"# URL: {url}\n[Blocked: URL is not allowed]"
+
             import urllib.request
             
             req = urllib.request.Request(

src/praisonai-agents/praisonaiagents/tools/python_tools.py

@@ -36,6 +36,30 @@ def _safe_getattr(obj, name, *default):
     return getattr(obj, name, *default) if default else getattr(obj, name)
 
 
+_SANDBOX_BLOCKED_ATTRS = frozenset({
+    '__subclasses__', '__bases__', '__mro__', '__globals__',
+    '__code__', '__class__', '__dict__', '__builtins__',
+    '__import__', '__loader__', '__spec__', '__init_subclass__',
+    '__set_name__', '__reduce__', '__reduce_ex__',
+    '__traceback__', '__qualname__', '__module__',
+    '__wrapped__', '__closure__', '__annotations__',
+    '__self__',  # C builtins leak real builtins module (GHSA-4mr5-g6f9-cfrh)
+    # Frame/code object introspection
+    'gi_frame', 'gi_code', 'cr_frame', 'cr_code',
+    'ag_frame', 'ag_code', 'tb_frame', 'tb_next',
+    'f_globals', 'f_locals', 'f_builtins', 'f_code',
+    'co_consts', 'co_names',
+    '__getattribute__', '__getattr__', '__setattr__', '__delattr__',
+    '__dir__', '__get__', '__set__', '__delete__',
+})
+
+_SANDBOX_BLOCKED_CALLS = frozenset({
+    'exec', 'eval', 'compile', '__import__',
+    'open', 'input', 'breakpoint',
+    'setattr', 'delattr', 'dir', 'vars',
+})
+
+
 def _validate_code_ast(code: str):
     """Validate code using AST — catches attacks that bypass text checks.
 
@@ -48,23 +72,6 @@ def _validate_code_ast(code: str):
     except SyntaxError:
         return None  # let compile() handle syntax errors later
 
-    # Dangerous dunder attributes attackers use for sandbox escape
-    _blocked_attrs = frozenset({
-        '__subclasses__', '__bases__', '__mro__', '__globals__',
-        '__code__', '__class__', '__dict__', '__builtins__',
-        '__import__', '__loader__', '__spec__', '__init_subclass__',
-        '__set_name__', '__reduce__', '__reduce_ex__',
-        '__traceback__', '__qualname__', '__module__',
-        '__wrapped__', '__closure__', '__annotations__',
-        # Frame/code object introspection
-        'gi_frame', 'gi_code', 'cr_frame', 'cr_code',
-        'ag_frame', 'ag_code', 'tb_frame', 'tb_next',
-        'f_globals', 'f_locals', 'f_builtins', 'f_code',
-        'co_consts', 'co_names',
-        '__getattribute__', '__getattr__', '__setattr__', '__delattr__',
-        '__dir__', '__get__', '__set__', '__delete__',
-    })
-
     for node in ast.walk(tree):
         # Block import statements
         if isinstance(node, (ast.Import, ast.ImportFrom)):
@@ -72,28 +79,26 @@ def _validate_code_ast(code: str):
 
         # Block attribute access to dangerous dunders
         if isinstance(node, ast.Attribute):
-            if node.attr in _blocked_attrs:
+            if node.attr in _SANDBOX_BLOCKED_ATTRS:
                 return (
                     f"Access to attribute '{node.attr}' is restricted"
                 )
 
-        # Block calls to dangerous builtins by name
+        # Block calls to dangerous builtins (bare name or attribute access)
         if isinstance(node, ast.Call):
             func = node.func
-            if isinstance(func, ast.Name) and func.id in (
-                'exec', 'eval', 'compile', '__import__',
-                'open', 'input', 'breakpoint',
-                'setattr', 'delattr', 'dir',
-            ):
+            if isinstance(func, ast.Name) and func.id in _SANDBOX_BLOCKED_CALLS:
                 return f"Call to '{func.id}' is not allowed"
+            if isinstance(func, ast.Attribute) and func.attr in _SANDBOX_BLOCKED_CALLS:
+                return f"Call to '{func.attr}' is not allowed"
                 
         # Block dangerous constants (strings containing dunders)
         # Fallback for Python 3.7 ast.Str
         if isinstance(node, ast.Constant) and isinstance(node.value, str):
-            if any(attr in node.value for attr in _blocked_attrs):
+            if any(attr in node.value for attr in _SANDBOX_BLOCKED_ATTRS):
                 return f"String constant contains restricted attribute name"
         elif type(node).__name__ == 'Str':
-            if any(attr in getattr(node, 's', '') for attr in _blocked_attrs):
+            if any(attr in getattr(node, 's', '') for attr in _SANDBOX_BLOCKED_ATTRS):
                 return f"String constant contains restricted attribute name"
 
     return None
@@ -112,7 +117,16 @@ def _execute_code_sandboxed(
     """
     if limits is None:
         limits = ResourceLimits.minimal()
-    
+
+    ast_error = _validate_code_ast(code)
+    if ast_error:
+        return {
+            'result': None,
+            'stdout': '',
+            'stderr': f'Security Error: {ast_error}',
+            'success': False,
+        }
+
     try:
         # Create temporary file for the code
         with tempfile.NamedTemporaryFile(mode='w', suffix='.py', delete=False) as f:
@@ -151,21 +165,8 @@ def safe_execute():
             }}
         
         # Block dangerous patterns
-        blocked_attrs = {{
-            '__subclasses__', '__bases__', '__mro__', '__globals__',
-            '__code__', '__class__', '__dict__', '__builtins__',
-            '__import__', '__loader__', '__spec__', '__init_subclass__',
-            '__set_name__', '__reduce__', '__reduce_ex__',
-            '__traceback__', '__qualname__', '__module__',
-            '__wrapped__', '__closure__', '__annotations__',
-            # Frame/code object introspection
-            'gi_frame', 'gi_code', 'cr_frame', 'cr_code',
-            'ag_frame', 'ag_code', 'tb_frame', 'tb_next',
-            'f_globals', 'f_locals', 'f_builtins', 'f_code',
-            'co_consts', 'co_names',
-            '__getattribute__', '__getattr__', '__setattr__', '__delattr__',
-            '__dir__', '__get__', '__set__', '__delete__',
-        }}
+        blocked_attrs = {set(_SANDBOX_BLOCKED_ATTRS)!r}
+        blocked_calls = {set(_SANDBOX_BLOCKED_CALLS)!r}
         
         for node in ast.walk(tree):
             if isinstance(node, (ast.Import, ast.ImportFrom)):
@@ -182,14 +183,20 @@ def safe_execute():
                     "stderr": f"Access to attribute '{{node.attr}}' is restricted",
                     "success": False
                 }}
-            if isinstance(node, ast.Call) and isinstance(node.func, ast.Name):
-                if node.func.id in ('exec', 'eval', 'compile', '__import__',
-                                     'open', 'input', 'breakpoint',
-                                     'setattr', 'delattr', 'dir'):
+            if isinstance(node, ast.Call):
+                func = node.func
+                if isinstance(func, ast.Name) and func.id in blocked_calls:
                     return {{
                         "result": None,
                         "stdout": "",
-                        "stderr": f"Call to '{{node.func.id}}' is not allowed",
+                        "stderr": f"Call to '{{func.id}}' is not allowed",
+                        "success": False
+                    }}
+                if isinstance(func, ast.Attribute) and func.attr in blocked_calls:
+                    return {{
+                        "result": None,
+                        "stdout": "",
+                        "stderr": f"Call to '{{func.attr}}' is not allowed",
                         "success": False
                     }}
             if isinstance(node, ast.Constant) and isinstance(node.value, str):
@@ -481,7 +488,7 @@ def _execute_code_direct(
             '__import__', 'import ', 'from ', 'exec', 'eval',
             'compile', 'open(', 'file(', 'input(', 'raw_input',
             '__subclasses__', '__bases__', '__globals__', '__code__',
-            '__class__', 'globals(', 'locals(', 'vars('
+            '__class__', '__self__', 'globals(', 'locals(', 'vars('
         ]
 
         code_lower = code.lower()

src/praisonai-agents/praisonaiagents/tools/registry.py

@@ -184,6 +184,41 @@ class ToolRegistry:
             
             return available
     
+    def list_tools_with_allowed_filter(self, context: Optional[Dict[str, Any]] = None) -> List[str]:
+        """List tool names filtered by ALLOWED_TOOLS environment variable.
+        
+        This applies the canonical ALLOWED_TOOLS filter to prevent tool
+        name collisions in multi-environment agent systems.
+        
+        Args:
+            context: Optional context for availability checks
+            
+        Returns:
+            List of filtered tool names
+        """
+        try:
+            # Lazy import to avoid circular dependencies
+            from ..allowed_tools_filter import filter_tools_with_allowed_tools
+            
+            # Get all available tool names
+            all_tools = self.list_tools()
+            
+            # Apply ALLOWED_TOOLS filter with diagnostics
+            filtered_tools = filter_tools_with_allowed_tools(all_tools, log_diagnostics=True)
+            
+            return sorted(filtered_tools)
+            
+        except ImportError:
+            logging.warning("ALLOWED_TOOLS filter not available, returning all tools")
+            return self.list_tools()
+        except ValueError:
+            # Preserve strict ALLOWED_TOOLS semantics (e.g., CI unknown tools, empty value)
+            raise
+        except Exception as e:
+            logging.error("Error applying ALLOWED_TOOLS filter: %s", e)
+            # Fallback to all tools on error
+            return self.list_tools()
+    
     def get_all(self) -> Dict[str, Union[BaseTool, Callable]]:
         """Get all registered tools as a dict. Thread-safe."""
         with self._lock:
@@ -381,3 +416,24 @@ def list_available_tools(context: Optional[Dict[str, Any]] = None) -> List[Union
         List of available tools
     """
     return get_registry().list_available_tools(context)
+
+
+def list_tools_with_allowed_filter(context: Optional[Dict[str, Any]] = None) -> List[str]:
+    """List tool names filtered by ALLOWED_TOOLS from the global registry.
+    
+    This is the canonical entry point for applying ALLOWED_TOOLS filtering
+    across the PraisonAI ecosystem.
+    
+    Args:
+        context: Optional context for availability checks
+        
+    Returns:
+        List of filtered tool names
+    """
+    return get_registry().list_tools_with_allowed_filter(context)
+
+
+# Backward compatibility alias
+def list_tools_with_hermes_filter(context: Optional[Dict[str, Any]] = None) -> List[str]:
+    """Backward compatibility alias for list_tools_with_allowed_filter."""
+    return list_tools_with_allowed_filter(context)

src/praisonai-agents/praisonaiagents/tools/spider_tools.py

@@ -11,6 +11,8 @@ content = scrape_page("https://example.com")
 """
 
 import logging
+import ipaddress
+import socket
 from typing import List, Dict, Union, Optional, Any
 from importlib import util
 import json
@@ -20,6 +22,48 @@ import os
 import hashlib
 import time
 
+
+def _host_is_blocked(hostname: str) -> bool:
+    """Return True when hostname resolves to loopback/private/internal targets."""
+    if not hostname:
+        return True
+    host = hostname.lower().rstrip(".")
+    if host in ("localhost", "0.0.0.0", "::1") or host.endswith(".localhost"):
+        return True
+    if host in ("169.254.169.254", "metadata.google.internal"):
+        return True
+    if any(host.endswith(suffix) for suffix in (".local", ".internal", ".localdomain")):
+        return True
+
+    def _ip_blocked(ip: ipaddress._BaseAddress) -> bool:
+        return bool(
+            ip.is_private or ip.is_loopback or ip.is_reserved or ip.is_link_local
+        )
+
+    if host.isdigit():
+        try:
+            return _ip_blocked(ipaddress.ip_address(int(host)))
+        except (ValueError, OverflowError):
+            return True
+
+    if host.startswith("0x"):
+        try:
+            return _ip_blocked(ipaddress.ip_address(int(host, 16)))
+        except (ValueError, OverflowError):
+            return True
+
+    try:
+        return _ip_blocked(ipaddress.ip_address(host))
+    except ValueError:
+        pass
+
+    try:
+        return _ip_blocked(ipaddress.ip_address(socket.inet_aton(host)))
+    except OSError:
+        pass
+
+    return False
+
 class SpiderTools:
     """Tools for web scraping and crawling."""
     
@@ -59,31 +103,9 @@ class SpiderTools:
             if not parsed.hostname:
                 return False
             
-            # Reject local/internal addresses
-            hostname = parsed.hostname.lower()
-            
-            # Block localhost and loopback
-            if hostname in ['localhost', '127.0.0.1', '0.0.0.0', '::1']:
+            if _host_is_blocked(parsed.hostname):
                 return False
-            
-            # Block private IP ranges
-            import ipaddress
-            try:
-                ip = ipaddress.ip_address(hostname)
-                if ip.is_private or ip.is_reserved or ip.is_loopback or ip.is_link_local:
-                    return False
-            except ValueError:
-                # Not an IP address, continue with domain validation
-                pass
-            
-            # Block common internal domains
-            if any(hostname.endswith(domain) for domain in ['.local', '.internal', '.localdomain']):
-                return False
-            
-            # Block metadata service endpoints
-            if hostname in ['169.254.169.254', 'metadata.google.internal']:
-                return False
-            
+
             return True
             
         except Exception:

src/praisonai-agents/praisonaiagents/ui/a2ui/__init__.py

@@ -1,126 +1,41 @@
 """
-A2UI (Agent-to-User Interface) Protocol Integration for PraisonAI Agents
+A2UI integration via Google's a2ui-agent-sdk (optional).
 
-This module provides A2UI protocol support, enabling PraisonAI Agents
-to generate rich, interactive UIs declaratively via JSON.
-
-A2UI is Google's open standard that allows agents to "speak UI" by sending
-declarative JSON describing UI components, which clients then render natively.
-
-Usage:
-    from praisonaiagents.ui.a2ui import (
-        # Message types
-        CreateSurfaceMessage,
-        UpdateComponentsMessage,
-        UpdateDataModelMessage,
-        DeleteSurfaceMessage,
-        # Component types
-        TextComponent,
-        ImageComponent,
-        ButtonComponent,
-        CardComponent,
-        RowComponent,
-        ColumnComponent,
-        # Extension helpers
-        create_a2ui_part,
-        is_a2ui_part,
-    )
+Declarative agent UI uses the official SDK — not reimplemented in PraisonAI core.
+Install: pip install praisonaiagents[a2ui]
 """
 
-from praisonaiagents.ui.a2ui.types import (
-    # Data binding
-    PathBinding,
-    resolve_string_or_path,
-    # Action types
-    ActionContext,
-    Action,
-    # Children types
-    ChildrenTemplate,
-    # Component types
-    TextComponent,
-    ImageComponent,
-    IconComponent,
-    ButtonComponent,
-    CardComponent,
-    RowComponent,
-    ColumnComponent,
-    ListComponent,
-    TextFieldComponent,
-    CheckBoxComponent,
-    SliderComponent,
-    DateTimeInputComponent,
-    # Message types
-    CreateSurfaceMessage,
-    UpdateComponentsMessage,
-    UpdateDataModelMessage,
-    DeleteSurfaceMessage,
-    # Data part
-    A2UIDataPart,
-)
-
-from praisonaiagents.ui.a2ui.extension import (
-    A2UI_MIME_TYPE,
-    A2UI_EXTENSION_URI,
-    STANDARD_CATALOG_ID,
+from praisonaiagents.ui.a2ui.adapter import (
     create_a2ui_part,
+    generate_a2ui_system_prompt,
+    get_schema_manager,
     is_a2ui_part,
-    get_a2ui_agent_extension,
-    AgentExtension,
-)
-
-from praisonaiagents.ui.a2ui.surface import Surface
-from praisonaiagents.ui.a2ui.agent import A2UIAgent
-from praisonaiagents.ui.a2ui.templates import (
-    ChatTemplate,
-    ListTemplate,
-    FormTemplate,
-    DashboardTemplate,
+    parse_a2ui_response,
 )
 
 __all__ = [
-    # Data binding
-    "PathBinding",
-    "resolve_string_or_path",
-    # Action types
-    "ActionContext",
-    "Action",
-    # Children types
-    "ChildrenTemplate",
-    # Component types
-    "TextComponent",
-    "ImageComponent",
-    "IconComponent",
-    "ButtonComponent",
-    "CardComponent",
-    "RowComponent",
-    "ColumnComponent",
-    "ListComponent",
-    "TextFieldComponent",
-    "CheckBoxComponent",
-    "SliderComponent",
-    "DateTimeInputComponent",
-    # Message types
-    "CreateSurfaceMessage",
-    "UpdateComponentsMessage",
-    "UpdateDataModelMessage",
-    "DeleteSurfaceMessage",
-    # Data part
-    "A2UIDataPart",
-    # Extension
-    "A2UI_MIME_TYPE",
-    "A2UI_EXTENSION_URI",
-    "STANDARD_CATALOG_ID",
+    "A2UI",
     "create_a2ui_part",
+    "generate_a2ui_system_prompt",
+    "get_schema_manager",
     "is_a2ui_part",
-    "get_a2ui_agent_extension",
-    "AgentExtension",
-    # Surface builder
-    "Surface",
-    # Agent wrapper
-    "A2UIAgent",
-    # Templates
-    "ChatTemplate",
-    "ListTemplate",
-    "FormTemplate",
-    "DashboardTemplate",
+    "parse_a2ui_response",
 ]
+
+
+class A2UI:
+    """
+    Optional A2UI facade (lazy — imports a2ui-agent-sdk on first use).
+
+    Example:
+        from praisonaiagents.ui import A2UI
+
+        part = A2UI.create_part({"createSurface": {...}})
+        prompt = A2UI.system_prompt("You are a helpful assistant.")
+    """
+
+    create_part = staticmethod(create_a2ui_part)
+    is_part = staticmethod(is_a2ui_part)
+    parse_response = staticmethod(parse_a2ui_response)
+    schema_manager = staticmethod(get_schema_manager)
+    system_prompt = staticmethod(generate_a2ui_system_prompt)

src/praisonai-agents/praisonaiagents/ui/a2ui/adapter.py

@@ -0,0 +1,89 @@
+"""
+Thin adapter over Google's a2ui-agent-sdk (optional dependency).
+
+Install: pip install praisonaiagents[a2ui]
+Spec: https://github.com/google/A2UI
+"""
+
+from __future__ import annotations
+
+from typing import Any, Dict, List, Optional
+
+
+_A2UI_INSTALL_HINT = (
+    "A2UI support requires the optional dependency. "
+    "Install with: pip install praisonaiagents[a2ui]"
+)
+
+
+def _import(name: str):
+    try:
+        import importlib
+
+        return importlib.import_module(name)
+    except ImportError as exc:
+        raise ImportError(
+            f"{_A2UI_INSTALL_HINT} "
+            f"(failed to import '{name}'; "
+            "if already installed, ensure a2a-sdk>=0.3.0,<1.0.0 and not a2a-sdk 1.x)"
+        ) from exc
+
+
+def create_a2ui_part(a2ui_data: Dict[str, Any]) -> Any:
+    """Wrap A2UI payload for A2A transport (MIME: application/json+a2ui)."""
+    parts = _import("a2ui.a2a.parts")
+    return parts.create_a2ui_part(a2ui_data)
+
+
+def is_a2ui_part(part: Any) -> bool:
+    """Return True if an A2A part contains A2UI data."""
+    parts = _import("a2ui.a2a.parts")
+    return parts.is_a2ui_part(part)
+
+
+def parse_a2ui_response(text: str) -> List[Any]:
+    """Parse and split LLM text into A2A parts (text + validated A2UI JSON)."""
+    parser = _import("a2ui.parser.parser")
+    return parser.parse_response(text)
+
+
+def get_schema_manager(
+    version: str = "0.9",
+    catalogs: Optional[List[Any]] = None,
+    accepts_inline_catalogs: bool = False,
+) -> Any:
+    """Create an A2uiSchemaManager for LLM system prompts and validation."""
+    constants = _import("a2ui.schema.constants")
+    manager = _import("a2ui.schema.manager")
+    basic = _import("a2ui.basic_catalog.provider")
+
+    version_const = constants.VERSION_0_9 if version == "0.9" else version
+    catalog_configs = catalogs
+    if catalog_configs is None:
+        catalog_configs = [basic.BasicCatalog.get_config(version=version_const)]
+
+    return manager.A2uiSchemaManager(
+        version=version_const,
+        catalogs=catalog_configs,
+        accepts_inline_catalogs=accepts_inline_catalogs,
+    )
+
+
+def generate_a2ui_system_prompt(
+    role_description: str,
+    workflow_description: str = "",
+    ui_description: str = "",
+    *,
+    version: str = "0.9",
+    include_schema: bool = True,
+    include_examples: bool = True,
+) -> str:
+    """Build an LLM system prompt with embedded A2UI schema and examples."""
+    schema_manager = get_schema_manager(version=version)
+    return schema_manager.generate_system_prompt(
+        role_description=role_description,
+        workflow_description=workflow_description,
+        ui_description=ui_description,
+        include_schema=include_schema,
+        include_examples=include_examples,
+    )

src/praisonai-agents/praisonaiagents/ui/a2ui/agent.py

@@ -1,358 +0,0 @@
-"""
-A2UI Agent Integration
-
-Provides A2UIAgent wrapper for generating A2UI responses from PraisonAI agents.
-"""
-
-import uuid
-from typing import Any, Dict, List, Optional, TYPE_CHECKING
-
-from praisonaiagents.ui.a2ui.surface import Surface
-from praisonaiagents.ui.a2ui.types import (
-    PathBinding,
-    A2UIDataPart,
-)
-from praisonaiagents.ui.a2ui.extension import (
-    create_a2ui_part,
-    STANDARD_CATALOG_ID,
-)
-
-if TYPE_CHECKING:
-    from praisonaiagents import Agent
-
-
-class A2UIAgent:
-    """
-    A2UI wrapper for PraisonAI agents.
-    
-    Provides methods to generate A2UI responses from agent interactions.
-    
-    Example:
-        >>> from praisonaiagents import Agent
-        >>> from praisonaiagents.ui.a2ui import A2UIAgent
-        >>> 
-        >>> agent = Agent(name="Assistant", role="Helper", goal="Help users")
-        >>> a2ui_agent = A2UIAgent(agent=agent)
-        >>> 
-        >>> # Render a simple text response
-        >>> messages = a2ui_agent.render_text("Hello, World!")
-        >>> 
-        >>> # Render a list of items
-        >>> messages = a2ui_agent.render_list(
-        ...     title="Results",
-        ...     items=[{"name": "Item 1"}, {"name": "Item 2"}]
-        ... )
-    """
-    
-    def __init__(
-        self,
-        agent: "Agent",
-        surface_id: Optional[str] = None,
-        catalog_id: str = STANDARD_CATALOG_ID,
-    ):
-        """
-        Initialize A2UIAgent wrapper.
-        
-        Args:
-            agent: The PraisonAI Agent to wrap
-            surface_id: Unique ID for the UI surface (auto-generated if not provided)
-            catalog_id: The component catalog to use
-        """
-        self.agent = agent
-        self.surface_id = surface_id or f"surface-{uuid.uuid4().hex[:8]}"
-        self.catalog_id = catalog_id
-        self._surface: Optional[Surface] = None
-    
-    def get_surface(self) -> Surface:
-        """
-        Get the underlying Surface object.
-        
-        Creates a new Surface if one doesn't exist.
-        
-        Returns:
-            The Surface object
-        """
-        if self._surface is None:
-            self._surface = Surface(
-                surface_id=self.surface_id,
-                catalog_id=self.catalog_id,
-            )
-        return self._surface
-    
-    def _reset_surface(self) -> Surface:
-        """Reset and return a new Surface."""
-        self._surface = Surface(
-            surface_id=self.surface_id,
-            catalog_id=self.catalog_id,
-        )
-        return self._surface
-    
-    # =========================================================================
-    # Rendering methods
-    # =========================================================================
-    
-    def render_text(
-        self,
-        text: str,
-        title: Optional[str] = None,
-        usage_hint: str = "body",
-    ) -> List[Dict[str, Any]]:
-        """
-        Render a simple text response as A2UI.
-        
-        Args:
-            text: The text content to display
-            title: Optional title heading
-            usage_hint: Text style hint (h1, h2, body, caption)
-        
-        Returns:
-            List of A2UI message dictionaries
-        """
-        surface = self._reset_surface()
-        
-        children = []
-        
-        if title:
-            surface.text(id="title", text=title, usage_hint="h1")
-            children.append("title")
-        
-        surface.text(id="content", text=text, usage_hint=usage_hint)
-        children.append("content")
-        
-        # Add root column
-        surface.column(id="root", children=children)
-        
-        return surface.to_messages()
-    
-    def render_list(
-        self,
-        title: str,
-        items: List[Dict[str, Any]],
-        item_title_key: str = "title",
-        item_description_key: str = "description",
-        item_image_key: Optional[str] = None,
-        item_action: Optional[str] = None,
-    ) -> List[Dict[str, Any]]:
-        """
-        Render a list of items as A2UI cards.
-        
-        Args:
-            title: List title
-            items: List of item dictionaries
-            item_title_key: Key for item title in each dict
-            item_description_key: Key for item description
-            item_image_key: Optional key for item image URL
-            item_action: Optional action name for item buttons
-        
-        Returns:
-            List of A2UI message dictionaries
-        """
-        surface = self._reset_surface()
-        
-        # Title
-        surface.text(id="list-title", text=title, usage_hint="h1")
-        
-        # Create card for each item
-        card_ids = []
-        for i, item in enumerate(items):
-            card_id = f"card-{i}"
-            content_id = f"card-content-{i}"
-            title_id = f"item-title-{i}"
-            desc_id = f"item-desc-{i}"
-            
-            # Item title
-            surface.text(
-                id=title_id,
-                text=item.get(item_title_key, f"Item {i+1}"),
-                usage_hint="h3"
-            )
-            
-            # Item description
-            surface.text(
-                id=desc_id,
-                text=item.get(item_description_key, ""),
-                usage_hint="body"
-            )
-            
-            content_children = [title_id, desc_id]
-            
-            # Optional action button
-            if item_action:
-                btn_id = f"btn-{i}"
-                btn_text_id = f"btn-text-{i}"
-                surface.text(id=btn_text_id, text="Select")
-                surface.button(
-                    id=btn_id,
-                    child=btn_text_id,
-                    action_name=item_action,
-                    action_context=[{"key": "index", "value": i}],
-                    primary=True
-                )
-                content_children.append(btn_id)
-            
-            # Card content column
-            surface.column(id=content_id, children=content_children)
-            
-            # Card
-            surface.card(id=card_id, child=content_id)
-            card_ids.append(card_id)
-        
-        # List column
-        surface.column(id="item-list", children=card_ids)
-        
-        # Root column
-        surface.column(id="root", children=["list-title", "item-list"])
-        
-        # Set data
-        surface.set_data("items", items)
-        
-        return surface.to_messages()
-    
-    def render_form(
-        self,
-        title: str,
-        fields: List[Dict[str, str]],
-        submit_action: str = "submit",
-        submit_label: str = "Submit",
-    ) -> List[Dict[str, Any]]:
-        """
-        Render a form as A2UI.
-        
-        Args:
-            title: Form title
-            fields: List of field definitions with id, label, type
-            submit_action: Action name for submit button
-            submit_label: Label for submit button
-        
-        Returns:
-            List of A2UI message dictionaries
-        """
-        surface = self._reset_surface()
-        
-        # Title
-        surface.text(id="form-title", text=title, usage_hint="h1")
-        
-        # Fields
-        field_ids = []
-        form_data = {}
-        
-        for field in fields:
-            field_id = field.get("id", f"field-{len(field_ids)}")
-            label = field.get("label", field_id)
-            field_type = field.get("type", "text")
-            
-            surface.text_field(
-                id=field_id,
-                label=label,
-                text=PathBinding(path=f"/{field_id}"),
-                field_type=field_type,
-            )
-            field_ids.append(field_id)
-            form_data[field_id] = field.get("default", "")
-        
-        # Submit button
-        surface.text(id="submit-text", text=submit_label)
-        
-        # Build action context from fields
-        action_context = [
-            {"key": f["id"], "value": {"path": f"/{f['id']}"}}
-            for f in fields
-        ]
-        
-        surface.button(
-            id="submit-btn",
-            child="submit-text",
-            action_name=submit_action,
-            action_context=action_context,
-            primary=True,
-        )
-        field_ids.append("submit-btn")
-        
-        # Form column
-        surface.column(id="form-fields", children=field_ids)
-        
-        # Root column
-        surface.column(id="root", children=["form-title", "form-fields"])
-        
-        # Set initial form data
-        for key, value in form_data.items():
-            surface.set_data(key, value)
-        
-        return surface.to_messages()
-    
-    # =========================================================================
-    # Agent interaction methods
-    # =========================================================================
-    
-    async def chat_async(self, message: str) -> str:
-        """
-        Send a message to the agent and get a response.
-        
-        Args:
-            message: The message to send
-        
-        Returns:
-            The agent's response text
-        """
-        # Use the agent's chat method
-        if hasattr(self.agent, 'chat'):
-            response = self.agent.chat(message)
-            return response
-        elif hasattr(self.agent, 'run'):
-            response = self.agent.run(message)
-            return response
-        else:
-            raise NotImplementedError("Agent does not have chat or run method")
-    
-    def chat(self, message: str) -> str:
-        """
-        Send a message to the agent and get a response (sync).
-        
-        Args:
-            message: The message to send
-        
-        Returns:
-            The agent's response text
-        """
-        if hasattr(self.agent, 'chat'):
-            return self.agent.chat(message)
-        elif hasattr(self.agent, 'run'):
-            return self.agent.run(message)
-        else:
-            raise NotImplementedError("Agent does not have chat or run method")
-    
-    # =========================================================================
-    # Export methods
-    # =========================================================================
-    
-    def to_messages(self) -> List[Dict[str, Any]]:
-        """
-        Get the current surface as A2UI messages.
-        
-        Returns:
-            List of A2UI message dictionaries
-        """
-        return self.get_surface().to_messages()
-    
-    def to_json(self, indent: Optional[int] = None) -> str:
-        """
-        Get the current surface as JSON string.
-        
-        Args:
-            indent: JSON indentation level
-        
-        Returns:
-            JSON string of A2UI messages
-        """
-        return self.get_surface().to_json(indent=indent)
-    
-    def to_a2a_part(self) -> A2UIDataPart:
-        """
-        Get the current surface as an A2A DataPart.
-        
-        Returns:
-            A2UIDataPart with A2UI MIME type
-        """
-        messages = self.to_messages()
-        # For A2A, we send messages as a list
-        return create_a2ui_part({"messages": messages})

src/praisonai-agents/praisonaiagents/ui/a2ui/extension.py

@@ -1,115 +0,0 @@
-"""
-A2UI Extension for A2A Protocol
-
-Provides helpers for integrating A2UI with the A2A (Agent-to-Agent) protocol.
-"""
-
-from dataclasses import dataclass
-from typing import Any, Dict, Optional
-
-from praisonaiagents.ui.a2ui.types import A2UIDataPart
-
-
-# =============================================================================
-# Constants
-# =============================================================================
-
-A2UI_EXTENSION_URI = "https://a2ui.org/a2a-extension/a2ui/v0.9"
-A2UI_MIME_TYPE = "application/json+a2ui"
-STANDARD_CATALOG_ID = "https://raw.githubusercontent.com/google/A2UI/refs/heads/main/specification/0.9/json/standard_catalog_definition.json"
-
-
-# =============================================================================
-# Agent Extension Type
-# =============================================================================
-
-@dataclass
-class AgentExtension:
-    """
-    A2A Agent Extension configuration.
-    
-    Describes an extension that an agent supports.
-    """
-    uri: str
-    description: str = ""
-    params: Optional[Dict[str, Any]] = None
-
-
-# =============================================================================
-# Helper Functions
-# =============================================================================
-
-def create_a2ui_part(a2ui_data: Dict[str, Any]) -> A2UIDataPart:
-    """
-    Create an A2UI data part for A2A integration.
-    
-    Wraps A2UI data in an A2UIDataPart with the appropriate MIME type.
-    
-    Args:
-        a2ui_data: The A2UI data dictionary (e.g., createSurface, updateComponents)
-    
-    Returns:
-        An A2UIDataPart with the A2UI MIME type in metadata
-    
-    Example:
-        >>> part = create_a2ui_part({"createSurface": {"surfaceId": "main", "catalogId": "..."}})
-        >>> part.metadata["mimeType"]
-        'application/json+a2ui'
-    """
-    return A2UIDataPart(
-        data=a2ui_data,
-        metadata={"mimeType": A2UI_MIME_TYPE}
-    )
-
-
-def is_a2ui_part(part: A2UIDataPart) -> bool:
-    """
-    Check if a data part contains A2UI data.
-    
-    Args:
-        part: The data part to check
-    
-    Returns:
-        True if the part contains A2UI data (has A2UI MIME type), False otherwise
-    
-    Example:
-        >>> part = create_a2ui_part({"createSurface": {...}})
-        >>> is_a2ui_part(part)
-        True
-    """
-    if not isinstance(part, A2UIDataPart):
-        return False
-    if part.metadata is None:
-        return False
-    return part.metadata.get("mimeType") == A2UI_MIME_TYPE
-
-
-def get_a2ui_agent_extension(
-    accepts_inline_custom_catalog: bool = False,
-) -> AgentExtension:
-    """
-    Create the A2UI AgentExtension configuration.
-    
-    Returns an AgentExtension that can be added to an A2A Agent Card
-    to indicate A2UI support.
-    
-    Args:
-        accepts_inline_custom_catalog: Whether the agent accepts inline custom catalogs
-    
-    Returns:
-        The configured A2UI AgentExtension
-    
-    Example:
-        >>> ext = get_a2ui_agent_extension()
-        >>> ext.uri
-        'https://a2ui.org/a2a-extension/a2ui/v0.9'
-    """
-    params: Optional[Dict[str, Any]] = None
-    if accepts_inline_custom_catalog:
-        params = {"acceptsInlineCustomCatalog": True}
-    
-    return AgentExtension(
-        uri=A2UI_EXTENSION_URI,
-        description="Provides agent-driven UI using the A2UI JSON format.",
-        params=params
-    )

src/praisonai-agents/praisonaiagents/ui/a2ui/surface.py

@@ -1,364 +0,0 @@
-"""
-A2UI Surface Builder
-
-Provides a fluent API for building A2UI surfaces with components and data.
-"""
-
-import json
-from typing import Any, Dict, List, Optional
-
-from praisonaiagents.ui.a2ui.extension import STANDARD_CATALOG_ID
-from praisonaiagents.ui.a2ui.types import (
-    Component,
-    TextComponent,
-    ImageComponent,
-    ButtonComponent,
-    CardComponent,
-    RowComponent,
-    ColumnComponent,
-    ListComponent,
-    TextFieldComponent,
-    Action,
-    ActionContext,
-    PathBinding,
-    CreateSurfaceMessage,
-    UpdateComponentsMessage,
-    UpdateDataModelMessage,
-    StringOrPath,
-    ChildrenType,
-)
-
-
-class Surface:
-    """
-    A2UI Surface builder for creating rich agent-generated UIs.
-    
-    Provides a fluent API for adding components and data, then
-    generating the A2UI JSON messages.
-    
-    Example:
-        >>> surface = Surface(surface_id="main")
-        >>> surface.text(id="title", text="Hello World", usage_hint="h1")
-        >>> surface.button(id="submit", child="submit-text", action_name="submit")
-        >>> surface.set_data("title", "Hello World")
-        >>> messages = surface.to_messages()
-    """
-    
-    def __init__(
-        self,
-        surface_id: str,
-        catalog_id: str = STANDARD_CATALOG_ID,
-    ):
-        """
-        Initialize a new Surface.
-        
-        Args:
-            surface_id: Unique identifier for this surface
-            catalog_id: The component catalog to use (defaults to standard catalog)
-        """
-        self.surface_id = surface_id
-        self.catalog_id = catalog_id
-        self.components: List[Component] = []
-        self.data: Dict[str, Any] = {}
-    
-    def add(self, component: Component) -> "Surface":
-        """
-        Add a component to the surface.
-        
-        Args:
-            component: The component to add
-        
-        Returns:
-            self for method chaining
-        """
-        self.components.append(component)
-        return self
-    
-    def set_data(self, key: str, value: Any) -> "Surface":
-        """
-        Set a value in the data model.
-        
-        Args:
-            key: The key in the data model
-            value: The value to set
-        
-        Returns:
-            self for method chaining
-        """
-        self.data[key] = value
-        return self
-    
-    # =========================================================================
-    # Helper methods for common components
-    # =========================================================================
-    
-    def text(
-        self,
-        id: str,
-        text: StringOrPath,
-        usage_hint: Optional[str] = None,
-        weight: Optional[float] = None,
-    ) -> "Surface":
-        """
-        Add a Text component.
-        
-        Args:
-            id: Component ID
-            text: Text content or path binding
-            usage_hint: Style hint (h1, h2, h3, body, caption)
-            weight: Flex weight
-        
-        Returns:
-            self for method chaining
-        """
-        return self.add(TextComponent(
-            id=id,
-            text=text,
-            usage_hint=usage_hint,
-            weight=weight,
-        ))
-    
-    def image(
-        self,
-        id: str,
-        url: StringOrPath,
-        fit: Optional[str] = None,
-        usage_hint: Optional[str] = None,
-        weight: Optional[float] = None,
-    ) -> "Surface":
-        """
-        Add an Image component.
-        
-        Args:
-            id: Component ID
-            url: Image URL or path binding
-            fit: Image fit (contain, cover, fill, none, scale-down)
-            usage_hint: Size hint (icon, avatar, smallFeature, etc.)
-            weight: Flex weight
-        
-        Returns:
-            self for method chaining
-        """
-        return self.add(ImageComponent(
-            id=id,
-            url=url,
-            fit=fit,
-            usage_hint=usage_hint,
-            weight=weight,
-        ))
-    
-    def button(
-        self,
-        id: str,
-        child: str,
-        action_name: Optional[str] = None,
-        action_context: Optional[List[Dict[str, Any]]] = None,
-        primary: bool = False,
-        weight: Optional[float] = None,
-    ) -> "Surface":
-        """
-        Add a Button component.
-        
-        Args:
-            id: Component ID
-            child: Child component ID (usually a Text component)
-            action_name: Name of the action to trigger
-            action_context: List of {key, value} dicts for action context
-            primary: Whether this is a primary button
-            weight: Flex weight
-        
-        Returns:
-            self for method chaining
-        """
-        action = None
-        if action_name:
-            context = []
-            if action_context:
-                for ctx in action_context:
-                    value = ctx.get("value")
-                    if isinstance(value, dict) and "path" in value:
-                        value = PathBinding(path=value["path"])
-                    context.append(ActionContext(key=ctx["key"], value=value))
-            action = Action(name=action_name, context=context)
-        
-        return self.add(ButtonComponent(
-            id=id,
-            child=child,
-            action=action,
-            primary=primary,
-            weight=weight,
-        ))
-    
-    def card(
-        self,
-        id: str,
-        child: str,
-        weight: Optional[float] = None,
-    ) -> "Surface":
-        """
-        Add a Card component.
-        
-        Args:
-            id: Component ID
-            child: Child component ID
-            weight: Flex weight
-        
-        Returns:
-            self for method chaining
-        """
-        return self.add(CardComponent(
-            id=id,
-            child=child,
-            weight=weight,
-        ))
-    
-    def row(
-        self,
-        id: str,
-        children: ChildrenType,
-        weight: Optional[float] = None,
-    ) -> "Surface":
-        """
-        Add a Row (horizontal layout) component.
-        
-        Args:
-            id: Component ID
-            children: List of child IDs or ChildrenTemplate
-            weight: Flex weight
-        
-        Returns:
-            self for method chaining
-        """
-        return self.add(RowComponent(
-            id=id,
-            children=children,
-            weight=weight,
-        ))
-    
-    def column(
-        self,
-        id: str,
-        children: ChildrenType,
-        weight: Optional[float] = None,
-    ) -> "Surface":
-        """
-        Add a Column (vertical layout) component.
-        
-        Args:
-            id: Component ID
-            children: List of child IDs or ChildrenTemplate
-            weight: Flex weight
-        
-        Returns:
-            self for method chaining
-        """
-        return self.add(ColumnComponent(
-            id=id,
-            children=children,
-            weight=weight,
-        ))
-    
-    def list(
-        self,
-        id: str,
-        children: ChildrenType,
-        direction: str = "vertical",
-        weight: Optional[float] = None,
-    ) -> "Surface":
-        """
-        Add a List component.
-        
-        Args:
-            id: Component ID
-            children: List of child IDs or ChildrenTemplate for dynamic lists
-            direction: "vertical" or "horizontal"
-            weight: Flex weight
-        
-        Returns:
-            self for method chaining
-        """
-        return self.add(ListComponent(
-            id=id,
-            children=children,
-            direction=direction,
-            weight=weight,
-        ))
-    
-    def text_field(
-        self,
-        id: str,
-        label: StringOrPath,
-        text: StringOrPath,
-        field_type: str = "text",
-        weight: Optional[float] = None,
-    ) -> "Surface":
-        """
-        Add a TextField input component.
-        
-        Args:
-            id: Component ID
-            label: Field label
-            text: Text value or path binding
-            field_type: Input type (text, number, email, etc.)
-            weight: Flex weight
-        
-        Returns:
-            self for method chaining
-        """
-        return self.add(TextFieldComponent(
-            id=id,
-            label=label,
-            text=text,
-            field_type=field_type,
-            weight=weight,
-        ))
-    
-    # =========================================================================
-    # Message generation
-    # =========================================================================
-    
-    def to_messages(self) -> List[Dict[str, Any]]:
-        """
-        Generate A2UI messages for this surface.
-        
-        Returns:
-            List of A2UI message dictionaries
-        """
-        messages: List[Dict[str, Any]] = []
-        
-        # 1. CreateSurface message
-        create_msg = CreateSurfaceMessage(
-            surface_id=self.surface_id,
-            catalog_id=self.catalog_id,
-        )
-        messages.append(create_msg.to_dict())
-        
-        # 2. UpdateComponents message (if we have components)
-        if self.components:
-            update_msg = UpdateComponentsMessage(
-                surface_id=self.surface_id,
-                components=self.components,
-            )
-            messages.append(update_msg.to_dict())
-        
-        # 3. UpdateDataModel message (if we have data)
-        if self.data:
-            data_msg = UpdateDataModelMessage(
-                surface_id=self.surface_id,
-                value=self.data,
-            )
-            messages.append(data_msg.to_dict())
-        
-        return messages
-    
-    def to_json(self, indent: Optional[int] = None) -> str:
-        """
-        Generate A2UI messages as a JSON string.
-        
-        Args:
-            indent: JSON indentation level (None for compact)
-        
-        Returns:
-            JSON string of A2UI messages
-        """
-        return json.dumps(self.to_messages(), indent=indent)

src/praisonai-agents/praisonaiagents/ui/a2ui/templates/__init__.py

@@ -1,17 +0,0 @@
-"""
-A2UI Templates
-
-Pre-built UI templates for common use cases.
-"""
-
-from praisonaiagents.ui.a2ui.templates.chat import ChatTemplate
-from praisonaiagents.ui.a2ui.templates.list import ListTemplate
-from praisonaiagents.ui.a2ui.templates.form import FormTemplate
-from praisonaiagents.ui.a2ui.templates.dashboard import DashboardTemplate
-
-__all__ = [
-    "ChatTemplate",
-    "ListTemplate",
-    "FormTemplate",
-    "DashboardTemplate",
-]

src/praisonai-agents/praisonaiagents/ui/a2ui/templates/chat.py

@@ -1,122 +0,0 @@
-"""
-Chat Template for A2UI
-
-Provides a conversational chat UI template.
-"""
-
-from typing import Any, Dict, List, Optional
-
-from praisonaiagents.ui.a2ui.surface import Surface
-from praisonaiagents.ui.a2ui.extension import STANDARD_CATALOG_ID
-
-
-class ChatTemplate:
-    """
-    Chat UI template for conversational interfaces.
-    
-    Creates a chat-style UI with user and agent messages.
-    
-    Example:
-        >>> template = ChatTemplate(surface_id="chat")
-        >>> template.add_user_message("Hello!")
-        >>> template.add_agent_message("Hi there! How can I help?")
-        >>> messages = template.to_messages()
-    """
-    
-    def __init__(
-        self,
-        surface_id: str = "chat",
-        catalog_id: str = STANDARD_CATALOG_ID,
-        title: Optional[str] = None,
-    ):
-        """
-        Initialize ChatTemplate.
-        
-        Args:
-            surface_id: Unique ID for the surface
-            catalog_id: Component catalog to use
-            title: Optional title for the chat
-        """
-        self.surface_id = surface_id
-        self.catalog_id = catalog_id
-        self.title = title
-        self.messages: List[Dict[str, Any]] = []
-    
-    def add_user_message(self, content: str) -> "ChatTemplate":
-        """
-        Add a user message.
-        
-        Args:
-            content: Message content
-        
-        Returns:
-            self for chaining
-        """
-        self.messages.append({
-            "role": "user",
-            "content": content,
-        })
-        return self
-    
-    def add_agent_message(self, content: str) -> "ChatTemplate":
-        """
-        Add an agent message.
-        
-        Args:
-            content: Message content
-        
-        Returns:
-            self for chaining
-        """
-        self.messages.append({
-            "role": "agent",
-            "content": content,
-        })
-        return self
-    
-    def to_messages(self) -> List[Dict[str, Any]]:
-        """
-        Generate A2UI messages for the chat.
-        
-        Returns:
-            List of A2UI message dictionaries
-        """
-        surface = Surface(
-            surface_id=self.surface_id,
-            catalog_id=self.catalog_id,
-        )
-        
-        children = []
-        
-        # Title if provided
-        if self.title:
-            surface.text(id="chat-title", text=self.title, usage_hint="h1")
-            children.append("chat-title")
-        
-        # Message components
-        msg_ids = []
-        for i, msg in enumerate(self.messages):
-            msg_id = f"msg-{i}"
-            role = msg["role"]
-            content = msg["content"]
-            
-            # Style based on role
-            if role == "user":
-                surface.text(id=msg_id, text=f"You: {content}", usage_hint="body")
-            else:
-                surface.text(id=msg_id, text=f"Agent: {content}", usage_hint="body")
-            
-            msg_ids.append(msg_id)
-        
-        # Messages column
-        if msg_ids:
-            surface.column(id="messages", children=msg_ids)
-            children.append("messages")
-        
-        # Root column
-        surface.column(id="root", children=children)
-        
-        # Set data
-        surface.set_data("messages", self.messages)
-        
-        return surface.to_messages()

src/praisonai-agents/praisonaiagents/ui/a2ui/templates/dashboard.py

@@ -1,128 +0,0 @@
-"""
-Dashboard Template for A2UI
-
-Provides a dashboard UI template with multiple panels.
-"""
-
-from typing import Any, Dict, List, Optional
-
-from praisonaiagents.ui.a2ui.surface import Surface
-from praisonaiagents.ui.a2ui.extension import STANDARD_CATALOG_ID
-
-
-class DashboardTemplate:
-    """
-    Dashboard UI template with multiple panels.
-    
-    Creates a dashboard layout with cards for each panel.
-    
-    Example:
-        >>> template = DashboardTemplate(surface_id="dashboard", title="My Dashboard")
-        >>> template.add_panel(id="stats", title="Statistics", content="100 items")
-        >>> template.add_panel(id="chart", title="Chart", content="[Chart here]")
-        >>> messages = template.to_messages()
-    """
-    
-    def __init__(
-        self,
-        surface_id: str = "dashboard",
-        title: str = "Dashboard",
-        catalog_id: str = STANDARD_CATALOG_ID,
-        columns: int = 2,
-    ):
-        """
-        Initialize DashboardTemplate.
-        
-        Args:
-            surface_id: Unique ID for the surface
-            title: Dashboard title
-            catalog_id: Component catalog to use
-            columns: Number of columns in the grid
-        """
-        self.surface_id = surface_id
-        self.title = title
-        self.catalog_id = catalog_id
-        self.columns = columns
-        self.panels: List[Dict[str, Any]] = []
-    
-    def add_panel(
-        self,
-        id: str,
-        title: str,
-        content: str,
-        weight: Optional[float] = None,
-    ) -> "DashboardTemplate":
-        """
-        Add a panel to the dashboard.
-        
-        Args:
-            id: Panel ID
-            title: Panel title
-            content: Panel content text
-            weight: Optional flex weight
-        
-        Returns:
-            self for chaining
-        """
-        self.panels.append({
-            "id": id,
-            "title": title,
-            "content": content,
-            "weight": weight,
-        })
-        return self
-    
-    def to_messages(self) -> List[Dict[str, Any]]:
-        """
-        Generate A2UI messages for the dashboard.
-        
-        Returns:
-            List of A2UI message dictionaries
-        """
-        surface = Surface(
-            surface_id=self.surface_id,
-            catalog_id=self.catalog_id,
-        )
-        
-        # Title
-        surface.text(id="dashboard-title", text=self.title, usage_hint="h1")
-        
-        # Create cards for each panel
-        card_ids = []
-        for panel in self.panels:
-            panel_id = panel["id"]
-            content_id = f"{panel_id}-content"
-            title_id = f"{panel_id}-title"
-            text_id = f"{panel_id}-text"
-            
-            # Panel title
-            surface.text(id=title_id, text=panel["title"], usage_hint="h3")
-            
-            # Panel content
-            surface.text(id=text_id, text=panel["content"], usage_hint="body")
-            
-            # Panel content column
-            surface.column(id=content_id, children=[title_id, text_id])
-            
-            # Panel card
-            surface.card(id=panel_id, child=content_id, weight=panel.get("weight"))
-            card_ids.append(panel_id)
-        
-        # Arrange panels in rows based on columns setting
-        row_ids = []
-        for i in range(0, len(card_ids), self.columns):
-            row_cards = card_ids[i:i + self.columns]
-            row_id = f"row-{i // self.columns}"
-            surface.row(id=row_id, children=row_cards)
-            row_ids.append(row_id)
-        
-        # Panels container
-        surface.column(id="panels", children=row_ids)
-        
-        # Root column
-        surface.column(id="root", children=["dashboard-title", "panels"])
-        
-        # Set data
-        surface.set_data("panels", self.panels)
-        
-        return surface.to_messages()

src/praisonai-agents/praisonaiagents/ui/a2ui/templates/form.py

@@ -1,205 +0,0 @@
-"""
-Form Template for A2UI
-
-Provides a form UI template for data collection.
-"""
-
-from typing import Any, Dict, List
-
-from praisonaiagents.ui.a2ui.surface import Surface
-from praisonaiagents.ui.a2ui.types import PathBinding
-from praisonaiagents.ui.a2ui.extension import STANDARD_CATALOG_ID
-
-
-class FormTemplate:
-    """
-    Form UI template for data collection.
-    
-    Creates a form with various input fields and a submit button.
-    
-    Example:
-        >>> template = FormTemplate(surface_id="form", title="Contact")
-        >>> template.add_text_field(id="name", label="Name")
-        >>> template.add_email_field(id="email", label="Email")
-        >>> template.set_submit_action("submit_form", "Send")
-        >>> messages = template.to_messages()
-    """
-    
-    def __init__(
-        self,
-        surface_id: str = "form",
-        title: str = "Form",
-        catalog_id: str = STANDARD_CATALOG_ID,
-    ):
-        """
-        Initialize FormTemplate.
-        
-        Args:
-            surface_id: Unique ID for the surface
-            title: Form title
-            catalog_id: Component catalog to use
-        """
-        self.surface_id = surface_id
-        self.title = title
-        self.catalog_id = catalog_id
-        self.fields: List[Dict[str, Any]] = []
-        self.submit_action = "submit"
-        self.submit_label = "Submit"
-    
-    def add_text_field(
-        self,
-        id: str,
-        label: str,
-        default: str = "",
-        required: bool = False,
-    ) -> "FormTemplate":
-        """
-        Add a text input field.
-        
-        Args:
-            id: Field ID
-            label: Field label
-            default: Default value
-            required: Whether field is required
-        
-        Returns:
-            self for chaining
-        """
-        self.fields.append({
-            "id": id,
-            "label": label,
-            "type": "text",
-            "default": default,
-            "required": required,
-        })
-        return self
-    
-    def add_number_field(
-        self,
-        id: str,
-        label: str,
-        default: float = 0,
-        required: bool = False,
-    ) -> "FormTemplate":
-        """
-        Add a number input field.
-        
-        Args:
-            id: Field ID
-            label: Field label
-            default: Default value
-            required: Whether field is required
-        
-        Returns:
-            self for chaining
-        """
-        self.fields.append({
-            "id": id,
-            "label": label,
-            "type": "number",
-            "default": default,
-            "required": required,
-        })
-        return self
-    
-    def add_email_field(
-        self,
-        id: str,
-        label: str,
-        default: str = "",
-        required: bool = False,
-    ) -> "FormTemplate":
-        """
-        Add an email input field.
-        
-        Args:
-            id: Field ID
-            label: Field label
-            default: Default value
-            required: Whether field is required
-        
-        Returns:
-            self for chaining
-        """
-        self.fields.append({
-            "id": id,
-            "label": label,
-            "type": "email",
-            "default": default,
-            "required": required,
-        })
-        return self
-    
-    def set_submit_action(
-        self,
-        action: str,
-        label: str = "Submit",
-    ) -> "FormTemplate":
-        """
-        Set the submit button action.
-        
-        Args:
-            action: Action name
-            label: Button label
-        
-        Returns:
-            self for chaining
-        """
-        self.submit_action = action
-        self.submit_label = label
-        return self
-    
-    def to_messages(self) -> List[Dict[str, Any]]:
-        """
-        Generate A2UI messages for the form.
-        
-        Returns:
-            List of A2UI message dictionaries
-        """
-        surface = Surface(
-            surface_id=self.surface_id,
-            catalog_id=self.catalog_id,
-        )
-        
-        # Title
-        surface.text(id="form-title", text=self.title, usage_hint="h1")
-        
-        # Fields
-        field_ids = []
-        for field in self.fields:
-            field_id = field["id"]
-            surface.text_field(
-                id=field_id,
-                label=field["label"],
-                text=PathBinding(path=f"/{field_id}"),
-                field_type=field["type"],
-            )
-            field_ids.append(field_id)
-            
-            # Set default value
-            surface.set_data(field_id, field.get("default", ""))
-        
-        # Submit button
-        surface.text(id="submit-text", text=self.submit_label)
-        
-        action_context = [
-            {"key": f["id"], "value": {"path": f"/{f['id']}"}}
-            for f in self.fields
-        ]
-        
-        surface.button(
-            id="submit-btn",
-            child="submit-text",
-            action_name=self.submit_action,
-            action_context=action_context,
-            primary=True,
-        )
-        field_ids.append("submit-btn")
-        
-        # Form fields column
-        surface.column(id="fields", children=field_ids)
-        
-        # Root column
-        surface.column(id="root", children=["form-title", "fields"])
-        
-        return surface.to_messages()

src/praisonai-agents/praisonaiagents/ui/a2ui/templates/list.py

@@ -1,147 +0,0 @@
-"""
-List Template for A2UI
-
-Provides a list UI template for displaying items.
-"""
-
-from typing import Any, Dict, List, Optional
-
-from praisonaiagents.ui.a2ui.surface import Surface
-from praisonaiagents.ui.a2ui.extension import STANDARD_CATALOG_ID
-
-
-class ListTemplate:
-    """
-    List UI template for displaying items.
-    
-    Creates a list of cards with title, description, and optional image.
-    
-    Example:
-        >>> template = ListTemplate(surface_id="list", title="Results")
-        >>> template.add_item(title="Item 1", description="First item")
-        >>> template.add_item(title="Item 2", description="Second item")
-        >>> messages = template.to_messages()
-    """
-    
-    def __init__(
-        self,
-        surface_id: str = "list",
-        title: str = "Items",
-        catalog_id: str = STANDARD_CATALOG_ID,
-    ):
-        """
-        Initialize ListTemplate.
-        
-        Args:
-            surface_id: Unique ID for the surface
-            title: List title
-            catalog_id: Component catalog to use
-        """
-        self.surface_id = surface_id
-        self.title = title
-        self.catalog_id = catalog_id
-        self.items: List[Dict[str, Any]] = []
-    
-    def add_item(
-        self,
-        title: str,
-        description: str = "",
-        image_url: Optional[str] = None,
-        action: Optional[str] = None,
-        metadata: Optional[Dict[str, Any]] = None,
-    ) -> "ListTemplate":
-        """
-        Add an item to the list.
-        
-        Args:
-            title: Item title
-            description: Item description
-            image_url: Optional image URL
-            action: Optional action name for item button
-            metadata: Optional additional metadata
-        
-        Returns:
-            self for chaining
-        """
-        item = {
-            "title": title,
-            "description": description,
-        }
-        if image_url:
-            item["image_url"] = image_url
-        if action:
-            item["action"] = action
-        if metadata:
-            item.update(metadata)
-        
-        self.items.append(item)
-        return self
-    
-    def to_messages(self) -> List[Dict[str, Any]]:
-        """
-        Generate A2UI messages for the list.
-        
-        Returns:
-            List of A2UI message dictionaries
-        """
-        surface = Surface(
-            surface_id=self.surface_id,
-            catalog_id=self.catalog_id,
-        )
-        
-        # Title
-        surface.text(id="list-title", text=self.title, usage_hint="h1")
-        
-        # Create cards for each item
-        card_ids = []
-        for i, item in enumerate(self.items):
-            card_id = f"card-{i}"
-            content_id = f"content-{i}"
-            title_id = f"title-{i}"
-            desc_id = f"desc-{i}"
-            
-            # Item title
-            surface.text(id=title_id, text=item["title"], usage_hint="h3")
-            
-            # Item description
-            surface.text(id=desc_id, text=item.get("description", ""), usage_hint="body")
-            
-            content_children = [title_id, desc_id]
-            
-            # Optional image
-            if item.get("image_url"):
-                img_id = f"img-{i}"
-                surface.image(id=img_id, url=item["image_url"], usage_hint="smallFeature")
-                content_children.insert(0, img_id)
-            
-            # Optional action button
-            if item.get("action"):
-                btn_id = f"btn-{i}"
-                btn_text_id = f"btn-text-{i}"
-                surface.text(id=btn_text_id, text="Select")
-                surface.button(
-                    id=btn_id,
-                    child=btn_text_id,
-                    action_name=item["action"],
-                    action_context=[{"key": "index", "value": i}],
-                    primary=True,
-                )
-                content_children.append(btn_id)
-            
-            # Card content
-            surface.column(id=content_id, children=content_children)
-            
-            # Card
-            surface.card(id=card_id, child=content_id)
-            card_ids.append(card_id)
-        
-        # Items column
-        surface.column(id="items", children=card_ids)
-        
-        # Root column
-        surface.column(id="root", children=["list-title", "items"])
-        
-        # Set data
-        surface.set_data("items", self.items)
-        
-        return surface.to_messages()

src/praisonai-agents/praisonaiagents/ui/a2ui/types.py

@@ -1,523 +0,0 @@
-"""
-A2UI Protocol Types
-
-Pydantic-style classes for the A2UI (Agent-to-User Interface) protocol.
-Based on the A2UI Specification v0.9.
-"""
-
-from dataclasses import dataclass, field
-from typing import Any, Dict, List, Optional, Union
-from enum import Enum
-
-
-# =============================================================================
-# Data Binding Types
-# =============================================================================
-
-@dataclass
-class PathBinding:
-    """
-    A reference to a value in the data model.
-    
-    Used for data binding in A2UI components.
-    """
-    path: str
-    
-    def to_dict(self) -> Dict[str, str]:
-        return {"path": self.path}
-
-
-# Type alias for values that can be literal or path-bound
-StringOrPath = Union[str, PathBinding]
-NumberOrPath = Union[int, float, PathBinding]
-BooleanOrPath = Union[bool, PathBinding]
-
-
-def resolve_string_or_path(value: StringOrPath) -> Union[str, Dict[str, str]]:
-    """Resolve a StringOrPath to its serialized form."""
-    if isinstance(value, PathBinding):
-        return value.to_dict()
-    return value
-
-
-def resolve_number_or_path(value: NumberOrPath) -> Union[int, float, Dict[str, str]]:
-    """Resolve a NumberOrPath to its serialized form."""
-    if isinstance(value, PathBinding):
-        return value.to_dict()
-    return value
-
-
-def resolve_boolean_or_path(value: BooleanOrPath) -> Union[bool, Dict[str, str]]:
-    """Resolve a BooleanOrPath to its serialized form."""
-    if isinstance(value, PathBinding):
-        return value.to_dict()
-    return value
-
-
-# =============================================================================
-# Action Types
-# =============================================================================
-
-@dataclass
-class ActionContext:
-    """
-    A key-value pair for action context.
-    
-    Used to pass data from UI to action handlers.
-    """
-    key: str
-    value: Union[str, int, float, bool, PathBinding]
-    
-    def to_dict(self) -> Dict[str, Any]:
-        if isinstance(self.value, PathBinding):
-            return {"key": self.key, "value": self.value.to_dict()}
-        return {"key": self.key, "value": self.value}
-
-
-@dataclass
-class Action:
-    """
-    An action triggered by a UI component (e.g., button click).
-    """
-    name: str
-    context: List[ActionContext] = field(default_factory=list)
-    
-    def to_dict(self) -> Dict[str, Any]:
-        return {
-            "name": self.name,
-            "context": [c.to_dict() for c in self.context]
-        }
-
-
-# =============================================================================
-# Children Types
-# =============================================================================
-
-@dataclass
-class ChildrenTemplate:
-    """
-    A template for generating dynamic children from a data model list.
-    """
-    component_id: str
-    path: str
-    
-    def to_dict(self) -> Dict[str, str]:
-        return {
-            "componentId": self.component_id,
-            "path": self.path
-        }
-
-
-# Type alias for children - either explicit list or template
-ChildrenType = Union[List[str], ChildrenTemplate]
-
-
-def resolve_children(children: ChildrenType) -> Union[List[str], Dict[str, str]]:
-    """Resolve children to serialized form."""
-    if isinstance(children, ChildrenTemplate):
-        return children.to_dict()
-    return children
-
-
-# =============================================================================
-# Component Types
-# =============================================================================
-
-class UsageHint(str, Enum):
-    """Text usage hints for styling."""
-    H1 = "h1"
-    H2 = "h2"
-    H3 = "h3"
-    H4 = "h4"
-    H5 = "h5"
-    CAPTION = "caption"
-    BODY = "body"
-
-
-class ImageFit(str, Enum):
-    """Image fit options."""
-    CONTAIN = "contain"
-    COVER = "cover"
-    FILL = "fill"
-    NONE = "none"
-    SCALE_DOWN = "scale-down"
-
-
-class ImageUsageHint(str, Enum):
-    """Image usage hints."""
-    ICON = "icon"
-    AVATAR = "avatar"
-    SMALL_FEATURE = "smallFeature"
-    MEDIUM_FEATURE = "mediumFeature"
-    LARGE_FEATURE = "largeFeature"
-    HEADER = "header"
-
-
-@dataclass
-class TextComponent:
-    """Text display component."""
-    id: str
-    text: StringOrPath
-    usage_hint: Optional[str] = None
-    weight: Optional[float] = None
-    
-    def to_dict(self) -> Dict[str, Any]:
-        result: Dict[str, Any] = {
-            "id": self.id,
-            "component": "Text",
-            "text": resolve_string_or_path(self.text)
-        }
-        if self.usage_hint:
-            result["usageHint"] = self.usage_hint
-        if self.weight is not None:
-            result["weight"] = self.weight
-        return result
-
-
-@dataclass
-class ImageComponent:
-    """Image display component."""
-    id: str
-    url: StringOrPath
-    fit: Optional[str] = None
-    usage_hint: Optional[str] = None
-    weight: Optional[float] = None
-    
-    def to_dict(self) -> Dict[str, Any]:
-        result: Dict[str, Any] = {
-            "id": self.id,
-            "component": "Image",
-            "url": resolve_string_or_path(self.url)
-        }
-        if self.fit:
-            result["fit"] = self.fit
-        if self.usage_hint:
-            result["usageHint"] = self.usage_hint
-        if self.weight is not None:
-            result["weight"] = self.weight
-        return result
-
-
-@dataclass
-class IconComponent:
-    """Icon display component."""
-    id: str
-    name: StringOrPath
-    weight: Optional[float] = None
-    
-    def to_dict(self) -> Dict[str, Any]:
-        result: Dict[str, Any] = {
-            "id": self.id,
-            "component": "Icon",
-            "name": resolve_string_or_path(self.name)
-        }
-        if self.weight is not None:
-            result["weight"] = self.weight
-        return result
-
-
-@dataclass
-class ButtonComponent:
-    """Button component with action."""
-    id: str
-    child: str
-    action: Optional[Action] = None
-    primary: bool = False
-    weight: Optional[float] = None
-    
-    def to_dict(self) -> Dict[str, Any]:
-        result: Dict[str, Any] = {
-            "id": self.id,
-            "component": "Button",
-            "child": self.child
-        }
-        if self.primary:
-            result["primary"] = True
-        if self.action:
-            result["action"] = self.action.to_dict()
-        if self.weight is not None:
-            result["weight"] = self.weight
-        return result
-
-
-@dataclass
-class CardComponent:
-    """Card container component."""
-    id: str
-    child: str
-    weight: Optional[float] = None
-    
-    def to_dict(self) -> Dict[str, Any]:
-        result: Dict[str, Any] = {
-            "id": self.id,
-            "component": "Card",
-            "child": self.child
-        }
-        if self.weight is not None:
-            result["weight"] = self.weight
-        return result
-
-
-@dataclass
-class RowComponent:
-    """Horizontal layout component."""
-    id: str
-    children: ChildrenType
-    weight: Optional[float] = None
-    
-    def to_dict(self) -> Dict[str, Any]:
-        result: Dict[str, Any] = {
-            "id": self.id,
-            "component": "Row",
-            "children": resolve_children(self.children)
-        }
-        if self.weight is not None:
-            result["weight"] = self.weight
-        return result
-
-
-@dataclass
-class ColumnComponent:
-    """Vertical layout component."""
-    id: str
-    children: ChildrenType
-    weight: Optional[float] = None
-    
-    def to_dict(self) -> Dict[str, Any]:
-        result: Dict[str, Any] = {
-            "id": self.id,
-            "component": "Column",
-            "children": resolve_children(self.children)
-        }
-        if self.weight is not None:
-            result["weight"] = self.weight
-        return result
-
-
-@dataclass
-class ListComponent:
-    """List component for repeating items."""
-    id: str
-    children: ChildrenType
-    direction: str = "vertical"
-    weight: Optional[float] = None
-    
-    def to_dict(self) -> Dict[str, Any]:
-        result: Dict[str, Any] = {
-            "id": self.id,
-            "component": "List",
-            "direction": self.direction,
-            "children": resolve_children(self.children)
-        }
-        if self.weight is not None:
-            result["weight"] = self.weight
-        return result
-
-
-@dataclass
-class TextFieldComponent:
-    """Text input field component."""
-    id: str
-    label: StringOrPath
-    text: StringOrPath
-    field_type: str = "text"
-    weight: Optional[float] = None
-    
-    def to_dict(self) -> Dict[str, Any]:
-        result: Dict[str, Any] = {
-            "id": self.id,
-            "component": "TextField",
-            "label": resolve_string_or_path(self.label),
-            "text": resolve_string_or_path(self.text),
-            "type": self.field_type
-        }
-        if self.weight is not None:
-            result["weight"] = self.weight
-        return result
-
-
-@dataclass
-class CheckBoxComponent:
-    """Checkbox input component."""
-    id: str
-    label: StringOrPath
-    checked: BooleanOrPath = False
-    weight: Optional[float] = None
-    
-    def to_dict(self) -> Dict[str, Any]:
-        result: Dict[str, Any] = {
-            "id": self.id,
-            "component": "CheckBox",
-            "label": resolve_string_or_path(self.label),
-            "checked": resolve_boolean_or_path(self.checked)
-        }
-        if self.weight is not None:
-            result["weight"] = self.weight
-        return result
-
-
-@dataclass
-class SliderComponent:
-    """Slider input component."""
-    id: str
-    min_value: NumberOrPath
-    max_value: NumberOrPath
-    value: NumberOrPath
-    weight: Optional[float] = None
-    
-    def to_dict(self) -> Dict[str, Any]:
-        result: Dict[str, Any] = {
-            "id": self.id,
-            "component": "Slider",
-            "min": resolve_number_or_path(self.min_value),
-            "max": resolve_number_or_path(self.max_value),
-            "value": resolve_number_or_path(self.value)
-        }
-        if self.weight is not None:
-            result["weight"] = self.weight
-        return result
-
-
-@dataclass
-class DateTimeInputComponent:
-    """Date/time input component."""
-    id: str
-    label: StringOrPath
-    value: StringOrPath
-    enable_date: bool = True
-    enable_time: bool = False
-    weight: Optional[float] = None
-    
-    def to_dict(self) -> Dict[str, Any]:
-        result: Dict[str, Any] = {
-            "id": self.id,
-            "component": "DateTimeInput",
-            "label": resolve_string_or_path(self.label),
-            "value": resolve_string_or_path(self.value),
-            "enableDate": self.enable_date,
-            "enableTime": self.enable_time
-        }
-        if self.weight is not None:
-            result["weight"] = self.weight
-        return result
-
-
-# Type alias for any component
-Component = Union[
-    TextComponent,
-    ImageComponent,
-    IconComponent,
-    ButtonComponent,
-    CardComponent,
-    RowComponent,
-    ColumnComponent,
-    ListComponent,
-    TextFieldComponent,
-    CheckBoxComponent,
-    SliderComponent,
-    DateTimeInputComponent,
-]
-
-
-# =============================================================================
-# Message Types
-# =============================================================================
-
-@dataclass
-class CreateSurfaceMessage:
-    """
-    Message to create a new UI surface.
-    
-    Signals the client to create a new surface and begin rendering it.
-    """
-    surface_id: str
-    catalog_id: str
-    
-    def to_dict(self) -> Dict[str, Any]:
-        return {
-            "createSurface": {
-                "surfaceId": self.surface_id,
-                "catalogId": self.catalog_id
-            }
-        }
-
-
-@dataclass
-class UpdateComponentsMessage:
-    """
-    Message to update surface components.
-    
-    Updates a surface with a new set of components.
-    """
-    surface_id: str
-    components: List[Component]
-    
-    def to_dict(self) -> Dict[str, Any]:
-        return {
-            "updateComponents": {
-                "surfaceId": self.surface_id,
-                "components": [c.to_dict() for c in self.components]
-            }
-        }
-
-
-@dataclass
-class UpdateDataModelMessage:
-    """
-    Message to update the data model.
-    
-    Updates the data model for an existing surface.
-    """
-    surface_id: str
-    value: Any
-    path: str = "/"
-    op: str = "replace"
-    
-    def to_dict(self) -> Dict[str, Any]:
-        result: Dict[str, Any] = {
-            "updateDataModel": {
-                "surfaceId": self.surface_id,
-                "path": self.path,
-                "op": self.op,
-                "value": self.value
-            }
-        }
-        return result
-
-
-@dataclass
-class DeleteSurfaceMessage:
-    """
-    Message to delete a surface.
-    
-    Signals the client to delete the surface.
-    """
-    surface_id: str
-    
-    def to_dict(self) -> Dict[str, Any]:
-        return {
-            "deleteSurface": {
-                "surfaceId": self.surface_id
-            }
-        }
-
-
-# =============================================================================
-# Data Part Type
-# =============================================================================
-
-@dataclass
-class A2UIDataPart:
-    """
-    A2UI data part for A2A integration.
-    
-    Wraps A2UI data with the appropriate MIME type.
-    """
-    data: Dict[str, Any]
-    metadata: Optional[Dict[str, Any]] = None
-    
-    def to_dict(self) -> Dict[str, Any]:
-        result: Dict[str, Any] = {"data": self.data}
-        if self.metadata:
-            result["metadata"] = self.metadata
-        return result

src/praisonai-agents/praisonaiagents/ui/agui/streaming.py

@@ -4,6 +4,8 @@ AG-UI Event Streaming
 Provides utilities for streaming PraisonAI agent responses as AG-UI events.
 """
 
+import asyncio
+import json
 import uuid
 from dataclasses import dataclass, field
 from typing import Any, AsyncIterator, Dict, Iterator, List, Optional, Set
@@ -256,6 +258,47 @@ async def async_stream_response(
         yield create_run_error_event(str(e))
 
 
+def stream_event_to_agui_events(
+    event: Any,
+    message_id: str,
+    buffer: EventBuffer,
+) -> List[BaseEvent]:
+    """Convert a PraisonAI StreamEvent into zero or more AG-UI events."""
+    from praisonaiagents.streaming.events import StreamEventType
+
+    if event.type == StreamEventType.DELTA_TEXT and event.content:
+        if not buffer.current_text_message_id:
+            buffer.start_text_message()
+            return [
+                TextMessageStartEvent(message_id=buffer.current_text_message_id, role="assistant"),
+                TextMessageContentEvent(
+                    message_id=buffer.current_text_message_id, delta=event.content
+                ),
+            ]
+        return [
+            TextMessageContentEvent(
+                message_id=buffer.current_text_message_id, delta=event.content
+            )
+        ]
+
+    if event.type == StreamEventType.TOOL_CALL_START and event.tool_call:
+        tool_call_id = event.tool_call.get("id") or str(uuid.uuid4())
+        tool_name = event.tool_call.get("name", "tool")
+        arguments = event.tool_call.get("arguments", {})
+        args_str = json.dumps(arguments) if isinstance(arguments, dict) else str(arguments)
+        parent_id = buffer.get_parent_message_id_for_tool_call() or message_id
+        buffer.start_tool_call(tool_call_id)
+        return list(create_tool_call_events(tool_call_id, tool_name, args_str, parent_id))
+
+    if event.type == StreamEventType.TOOL_CALL_RESULT and event.tool_call:
+        tool_call_id = event.tool_call.get("id") or str(uuid.uuid4())
+        content = event.content if event.content is not None else str(event.tool_call.get("result", ""))
+        buffer.end_tool_call(tool_call_id)
+        return [create_tool_result_event(tool_call_id, str(content), message_id)]
+
+    return []
+
+
 async def async_stream_agent_response(
     agent,
     user_input: str,
@@ -278,37 +321,63 @@ async def async_stream_agent_response(
     Yields:
         AG-UI events
     """
-    import asyncio
-    
-    # Emit run started
     yield create_run_started_event(thread_id, run_id)
-    
+
+    loop = asyncio.get_running_loop()
+    queue: asyncio.Queue = asyncio.Queue()
+    buffer = EventBuffer()
     message_id = str(uuid.uuid4())
-    
-    try:
-        # Check if agent has async chat method
-        if hasattr(agent, 'achat'):
-            response = await agent.achat(user_input)
-        elif hasattr(agent, 'chat'):
-            # Run sync chat in executor
-            loop = asyncio.get_event_loop()
-            response = await loop.run_in_executor(None, agent.chat, user_input)
-        else:
+    text_started = False
+    text_ended = False
+    original_emitter = getattr(agent, "stream_emitter", None)
+
+    def on_stream_event(event: Any) -> None:
+        loop.call_soon_threadsafe(queue.put_nowait, ("stream", event))
+
+    if original_emitter is not None:
+        original_emitter.add_callback(on_stream_event)
+
+    async def run_chat() -> Any:
+        try:
+            if hasattr(agent, "achat"):
+                return await agent.achat(user_input, stream=True)
+            if hasattr(agent, "chat"):
+                return await loop.run_in_executor(
+                    None, lambda: agent.chat(user_input, stream=True)
+                )
             raise ValueError("Agent must have 'chat' or 'achat' method")
-        
-        # Emit text message events
-        yield TextMessageStartEvent(message_id=message_id, role="assistant")
-        
-        if response:
+        finally:
+            loop.call_soon_threadsafe(queue.put_nowait, ("done", None))
+
+    chat_task = asyncio.create_task(run_chat())
+
+    try:
+        while True:
+            kind, payload = await queue.get()
+            if kind == "done":
+                break
+            for agui_event in stream_event_to_agui_events(payload, message_id, buffer):
+                if isinstance(agui_event, TextMessageStartEvent):
+                    text_started = True
+                yield agui_event
+
+        response = await chat_task
+
+        if text_started and buffer.current_text_message_id and not text_ended:
+            yield TextMessageEndEvent(message_id=buffer.current_text_message_id)
+            text_ended = True
+        elif not text_started and response:
+            yield TextMessageStartEvent(message_id=message_id, role="assistant")
             yield TextMessageContentEvent(message_id=message_id, delta=str(response))
-        
-        yield TextMessageEndEvent(message_id=message_id)
-        
-        # Emit run finished
+            yield TextMessageEndEvent(message_id=message_id)
+
         yield create_run_finished_event(thread_id, run_id)
-        
+
     except Exception as e:
         yield create_run_error_event(str(e))
+    finally:
+        if original_emitter is not None:
+            original_emitter.remove_callback(on_stream_event)
 
 
 async def async_stream_agents_response(
@@ -331,8 +400,6 @@ async def async_stream_agents_response(
     Yields:
         AG-UI events
     """
-    import asyncio
-    
     # Emit run started
     yield create_run_started_event(thread_id, run_id)
     
@@ -344,7 +411,7 @@ async def async_stream_agents_response(
             result = await agents.astart(user_input)
         elif hasattr(agents, 'start'):
             # Run sync start in executor
-            loop = asyncio.get_event_loop()
+            loop = asyncio.get_running_loop()
             result = await loop.run_in_executor(None, agents.start, user_input)
         else:
             raise ValueError("Agents must have 'start' or 'astart' method")

src/praisonai-agents/pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "setuptools.build_meta"
 
 [project]
 name = "praisonaiagents"
-version = "1.6.37"
+version = "1.6.44"
 description = "Praison AI agents for completing complex tasks with Self Reflection Agents"
 readme = "README.md"
 requires-python = ">=3.10"
@@ -101,6 +101,22 @@ crawl = [
     "playwright>=1.47.0",
 ]
 
+# Sandbox dependencies
+sandbox = [
+    "e2b-code-interpreter>=1.0.0",
+]
+
+# Declarative agent UI (Google A2UI SDK — lazy loaded)
+a2ui = [
+    "a2ui-agent-sdk>=0.2.1",
+    "a2a-sdk>=0.3.0,<1.0.0",
+]
+
+# Docker sandbox dependencies
+sandbox-docker = [
+    "docker>=7.0.0",
+]
+
 # Combined features
 all = [
     "praisonaiagents[memory]",
@@ -112,10 +128,12 @@ all = [
     "praisonaiagents[telemetry]",
     "praisonaiagents[mongodb]",
     "praisonaiagents[auth]",
+    "praisonaiagents[sandbox]",
     "praisonaiagents[search]",
     "praisonaiagents[crawl]",
     "praisonaiagents[autonomy]",
-    "praisonaiagents[os]"
+    "praisonaiagents[os]",
+    "praisonaiagents[a2ui]",
 ]
 
 [tool.setuptools.packages.find]

src/praisonai-agents/tests/a2ui/__init__.py

@@ -1 +0,0 @@
-"""A2UI Tests Package."""

src/praisonai-agents/tests/a2ui/test_adapter.py

@@ -0,0 +1,36 @@
+"""Tests for A2UI adapter (optional a2ui-agent-sdk dependency)."""
+
+import pytest
+
+
+class TestA2UIAdapterImport:
+    def test_a2ui_class_export(self):
+        from praisonaiagents.ui.a2ui import A2UI
+
+        assert A2UI is not None
+        assert hasattr(A2UI, "create_part")
+
+    def test_ui_lazy_a2ui_export(self):
+        from praisonaiagents.ui import A2UI
+
+        assert A2UI is not None
+
+    def test_missing_optional_dep_raises_clear_error(self):
+        pytest.importorskip("a2ui", reason="a2ui-agent-sdk not installed")
+        # If installed, adapter imports should work
+        from praisonaiagents.ui.a2ui.adapter import create_a2ui_part
+
+        part = create_a2ui_part({"createSurface": {"surfaceId": "main", "catalogId": "test"}})
+        assert part is not None
+
+
+class TestA2UITool:
+    def test_send_a2ui_messages_requires_sdk(self):
+        from praisonaiagents.tools.a2ui_tools import send_a2ui_messages
+
+        if pytest.importorskip("a2ui", reason="skip without sdk"):
+            result = send_a2ui_messages(
+                messages=[{"createSurface": {"surfaceId": "x", "catalogId": "c"}}]
+            )
+            assert result["mime_type"] == "application/json+a2ui"
+            assert len(result["messages"]) == 1

src/praisonai-agents/tests/a2ui/test_agent.py

@@ -1,202 +0,0 @@
-"""
-Tests for A2UI Agent Integration
-
-TDD: Write tests first, then implement the A2UIAgent class.
-"""
-
-import os
-import pytest
-
-
-# Skip tests if no API key
-pytestmark = pytest.mark.skipif(
-    not os.environ.get("OPENAI_API_KEY"),
-    reason="OPENAI_API_KEY not set"
-)
-
-
-class TestA2UIAgent:
-    """Test A2UIAgent wrapper class."""
-
-    def test_a2ui_agent_creation(self):
-        """Test A2UIAgent creation with an agent."""
-        from praisonaiagents import Agent
-        from praisonaiagents.ui.a2ui.agent import A2UIAgent
-        
-        agent = Agent(
-            name="TestAgent",
-            role="Test assistant",
-            goal="Help with testing"
-        )
-        
-        a2ui_agent = A2UIAgent(agent=agent)
-        
-        assert a2ui_agent.agent is agent
-        assert a2ui_agent.surface_id is not None
-
-    def test_a2ui_agent_with_custom_surface_id(self):
-        """Test A2UIAgent with custom surface ID."""
-        from praisonaiagents import Agent
-        from praisonaiagents.ui.a2ui.agent import A2UIAgent
-        
-        agent = Agent(name="TestAgent", role="Test", goal="Test")
-        a2ui_agent = A2UIAgent(agent=agent, surface_id="custom-surface")
-        
-        assert a2ui_agent.surface_id == "custom-surface"
-
-    def test_render_text_response(self):
-        """Test rendering a simple text response as A2UI."""
-        from praisonaiagents import Agent
-        from praisonaiagents.ui.a2ui.agent import A2UIAgent
-        
-        agent = Agent(name="TestAgent", role="Test", goal="Test")
-        a2ui_agent = A2UIAgent(agent=agent)
-        
-        # Render a simple text response
-        messages = a2ui_agent.render_text("Hello, World!")
-        
-        assert isinstance(messages, list)
-        assert len(messages) >= 2  # createSurface + updateComponents
-        
-        # Check createSurface
-        assert "createSurface" in messages[0]
-        
-        # Check updateComponents has a text component
-        update_msg = next((m for m in messages if "updateComponents" in m), None)
-        assert update_msg is not None
-        components = update_msg["updateComponents"]["components"]
-        assert any(c.get("component") == "Text" for c in components)
-
-    def test_render_list_response(self):
-        """Test rendering a list of items as A2UI cards."""
-        from praisonaiagents import Agent
-        from praisonaiagents.ui.a2ui.agent import A2UIAgent
-        
-        agent = Agent(name="TestAgent", role="Test", goal="Test")
-        a2ui_agent = A2UIAgent(agent=agent)
-        
-        items = [
-            {"title": "Item 1", "description": "First item"},
-            {"title": "Item 2", "description": "Second item"},
-        ]
-        
-        messages = a2ui_agent.render_list(
-            title="My Items",
-            items=items,
-            item_title_key="title",
-            item_description_key="description"
-        )
-        
-        assert isinstance(messages, list)
-        assert len(messages) >= 2
-
-    def test_render_form(self):
-        """Test rendering a form as A2UI."""
-        from praisonaiagents import Agent
-        from praisonaiagents.ui.a2ui.agent import A2UIAgent
-        
-        agent = Agent(name="TestAgent", role="Test", goal="Test")
-        a2ui_agent = A2UIAgent(agent=agent)
-        
-        fields = [
-            {"id": "name", "label": "Name", "type": "text"},
-            {"id": "email", "label": "Email", "type": "text"},
-        ]
-        
-        messages = a2ui_agent.render_form(
-            title="Contact Form",
-            fields=fields,
-            submit_action="submit_form"
-        )
-        
-        assert isinstance(messages, list)
-        assert len(messages) >= 2
-        
-        # Check for TextField components
-        update_msg = next((m for m in messages if "updateComponents" in m), None)
-        assert update_msg is not None
-        components = update_msg["updateComponents"]["components"]
-        text_fields = [c for c in components if c.get("component") == "TextField"]
-        assert len(text_fields) >= 2
-
-    def test_get_surface(self):
-        """Test getting the underlying Surface object."""
-        from praisonaiagents import Agent
-        from praisonaiagents.ui.a2ui.agent import A2UIAgent
-        from praisonaiagents.ui.a2ui.surface import Surface
-        
-        agent = Agent(name="TestAgent", role="Test", goal="Test")
-        a2ui_agent = A2UIAgent(agent=agent)
-        
-        surface = a2ui_agent.get_surface()
-        
-        assert isinstance(surface, Surface)
-        assert surface.surface_id == a2ui_agent.surface_id
-
-
-class TestA2UIAgentWithLLM:
-    """Test A2UIAgent with actual LLM calls."""
-
-    @pytest.mark.integration
-    @pytest.mark.asyncio
-    async def test_chat_with_ui(self):
-        """Test chat that returns A2UI response."""
-        from praisonaiagents import Agent
-        from praisonaiagents.ui.a2ui.agent import A2UIAgent
-        
-        agent = Agent(
-            name="UIAgent",
-            role="UI Generator",
-            goal="Generate helpful UI responses",
-            llm="gpt-4o-mini"
-        )
-        
-        a2ui_agent = A2UIAgent(agent=agent)
-        
-        # This should work with the agent's chat
-        response = await a2ui_agent.chat_async("Say hello")
-        
-        assert response is not None
-        # Response should contain text
-        assert isinstance(response, str) or isinstance(response, dict)
-
-
-class TestA2UIAgentIntegration:
-    """Integration tests for A2UIAgent with A2A."""
-
-    def test_get_a2a_data_part(self):
-        """Test getting A2UI as A2A DataPart."""
-        from praisonaiagents import Agent
-        from praisonaiagents.ui.a2ui.agent import A2UIAgent
-        from praisonaiagents.ui.a2ui.extension import is_a2ui_part
-        
-        agent = Agent(name="TestAgent", role="Test", goal="Test")
-        a2ui_agent = A2UIAgent(agent=agent)
-        
-        # Render something
-        a2ui_agent.render_text("Hello")
-        
-        # Get as A2A DataPart
-        part = a2ui_agent.to_a2a_part()
-        
-        assert part is not None
-        assert is_a2ui_part(part)
-
-    def test_to_json(self):
-        """Test getting A2UI as JSON string."""
-        from praisonaiagents import Agent
-        from praisonaiagents.ui.a2ui.agent import A2UIAgent
-        import json
-        
-        agent = Agent(name="TestAgent", role="Test", goal="Test")
-        a2ui_agent = A2UIAgent(agent=agent)
-        
-        # Render something
-        a2ui_agent.render_text("Hello")
-        
-        # Get as JSON
-        json_str = a2ui_agent.to_json()
-        
-        # Should be valid JSON
-        parsed = json.loads(json_str)
-        assert isinstance(parsed, list)

src/praisonai-agents/tests/a2ui/test_extension.py

@@ -1,115 +0,0 @@
-"""
-Tests for A2UI Extension Module
-
-TDD: Write tests first, then implement the extension.
-"""
-
-
-class TestA2UIExtension:
-    """Test A2UI extension helpers for A2A integration."""
-
-    def test_a2ui_mime_type(self):
-        """Test A2UI MIME type constant."""
-        from praisonaiagents.ui.a2ui.extension import A2UI_MIME_TYPE
-        
-        assert A2UI_MIME_TYPE == "application/json+a2ui"
-
-    def test_a2ui_extension_uri(self):
-        """Test A2UI extension URI constant."""
-        from praisonaiagents.ui.a2ui.extension import A2UI_EXTENSION_URI
-        
-        assert "a2ui" in A2UI_EXTENSION_URI.lower()
-
-    def test_create_a2ui_part(self):
-        """Test create_a2ui_part wraps A2UI data in A2A DataPart."""
-        from praisonaiagents.ui.a2ui.extension import create_a2ui_part, A2UI_MIME_TYPE
-        
-        a2ui_data = {
-            "createSurface": {
-                "surfaceId": "main",
-                "catalogId": "https://a2ui.dev/standard"
-            }
-        }
-        
-        part = create_a2ui_part(a2ui_data)
-        
-        # Should return a Part with DataPart root
-        assert part is not None
-        assert hasattr(part, 'data')
-        assert part.data == a2ui_data
-        assert part.metadata is not None
-        assert part.metadata.get("mimeType") == A2UI_MIME_TYPE
-
-    def test_is_a2ui_part_true(self):
-        """Test is_a2ui_part returns True for A2UI parts."""
-        from praisonaiagents.ui.a2ui.extension import create_a2ui_part, is_a2ui_part
-        
-        a2ui_data = {"createSurface": {"surfaceId": "main", "catalogId": "test"}}
-        part = create_a2ui_part(a2ui_data)
-        
-        assert is_a2ui_part(part) is True
-
-    def test_is_a2ui_part_false(self):
-        """Test is_a2ui_part returns False for non-A2UI parts."""
-        from praisonaiagents.ui.a2ui.extension import is_a2ui_part
-        from praisonaiagents.ui.a2ui.types import A2UIDataPart
-        
-        # Create a regular data part without A2UI mime type
-        part = A2UIDataPart(data={"foo": "bar"}, metadata={"mimeType": "application/json"})
-        
-        assert is_a2ui_part(part) is False
-
-    def test_get_a2ui_agent_extension(self):
-        """Test get_a2ui_agent_extension returns AgentExtension config."""
-        from praisonaiagents.ui.a2ui.extension import get_a2ui_agent_extension
-        
-        ext = get_a2ui_agent_extension()
-        
-        assert ext is not None
-        assert hasattr(ext, 'uri')
-        assert "a2ui" in ext.uri.lower()
-        assert hasattr(ext, 'description')
-
-    def test_get_a2ui_agent_extension_with_inline_catalog(self):
-        """Test get_a2ui_agent_extension with inline catalog support."""
-        from praisonaiagents.ui.a2ui.extension import get_a2ui_agent_extension
-        
-        ext = get_a2ui_agent_extension(accepts_inline_custom_catalog=True)
-        
-        assert ext.params is not None
-        assert ext.params.get("acceptsInlineCustomCatalog") is True
-
-    def test_standard_catalog_id(self):
-        """Test standard catalog ID constant."""
-        from praisonaiagents.ui.a2ui.extension import STANDARD_CATALOG_ID
-        
-        assert "standard_catalog" in STANDARD_CATALOG_ID.lower() or "a2ui" in STANDARD_CATALOG_ID.lower()
-
-
-class TestA2UIDataPart:
-    """Test A2UI DataPart type."""
-
-    def test_a2ui_data_part_creation(self):
-        """Test A2UIDataPart creation."""
-        from praisonaiagents.ui.a2ui.types import A2UIDataPart
-        
-        part = A2UIDataPart(
-            data={"test": "value"},
-            metadata={"mimeType": "application/json+a2ui"}
-        )
-        
-        assert part.data == {"test": "value"}
-        assert part.metadata["mimeType"] == "application/json+a2ui"
-
-    def test_a2ui_data_part_to_dict(self):
-        """Test A2UIDataPart serialization."""
-        from praisonaiagents.ui.a2ui.types import A2UIDataPart
-        
-        part = A2UIDataPart(
-            data={"createSurface": {"surfaceId": "main"}},
-            metadata={"mimeType": "application/json+a2ui"}
-        )
-        
-        data = part.to_dict()
-        assert "data" in data
-        assert "metadata" in data

src/praisonai-agents/tests/a2ui/test_integration.py

@@ -1,225 +0,0 @@
-"""
-End-to-End Integration Tests for A2UI
-
-Tests the complete A2UI workflow from agent to JSON output.
-"""
-
-import json
-import os
-import pytest
-
-
-class TestA2UIEndToEnd:
-    """End-to-end integration tests."""
-
-    def test_complete_surface_workflow(self):
-        """Test complete workflow: create surface, add components, generate JSON."""
-        from praisonaiagents.ui.a2ui import (
-            Surface,
-            PathBinding,
-        )
-        
-        # Create surface
-        surface = Surface(surface_id="test-surface")
-        
-        # Add components using fluent API
-        surface.text(id="title", text="Welcome", usage_hint="h1")
-        surface.text(id="subtitle", text=PathBinding(path="/subtitle"), usage_hint="h2")
-        surface.button(
-            id="action-btn",
-            child="btn-text",
-            action_name="do_action",
-            action_context=[{"key": "id", "value": "123"}],
-            primary=True
-        )
-        surface.text(id="btn-text", text="Click Me")
-        surface.column(id="root", children=["title", "subtitle", "action-btn"])
-        
-        # Set data
-        surface.set_data("subtitle", "Hello World")
-        
-        # Generate messages
-        messages = surface.to_messages()
-        
-        # Verify structure
-        assert len(messages) == 3  # createSurface, updateComponents, updateDataModel
-        
-        # Verify createSurface
-        assert "createSurface" in messages[0]
-        assert messages[0]["createSurface"]["surfaceId"] == "test-surface"
-        
-        # Verify updateComponents
-        assert "updateComponents" in messages[1]
-        components = messages[1]["updateComponents"]["components"]
-        assert len(components) == 5  # title, subtitle, btn, btn-text, root
-        
-        # Verify updateDataModel
-        assert "updateDataModel" in messages[2]
-        assert messages[2]["updateDataModel"]["value"]["subtitle"] == "Hello World"
-        
-        # Verify JSON output
-        json_str = surface.to_json()
-        parsed = json.loads(json_str)
-        assert len(parsed) == 3
-
-    def test_a2ui_agent_render_workflow(self):
-        """Test A2UIAgent rendering workflow."""
-        from praisonaiagents import Agent
-        from praisonaiagents.ui.a2ui import A2UIAgent
-        
-        agent = Agent(
-            name="TestAgent",
-            role="Test assistant",
-            goal="Help with testing"
-        )
-        
-        a2ui_agent = A2UIAgent(agent=agent, surface_id="agent-surface")
-        
-        # Render text
-        messages = a2ui_agent.render_text(
-            text="This is a test response",
-            title="Test Title"
-        )
-        
-        # Verify structure
-        assert len(messages) >= 2
-        assert "createSurface" in messages[0]
-        
-        # Verify JSON export
-        json_str = a2ui_agent.to_json()
-        parsed = json.loads(json_str)
-        assert isinstance(parsed, list)
-
-    def test_template_workflow(self):
-        """Test template-based workflow."""
-        from praisonaiagents.ui.a2ui import (
-            ChatTemplate,
-            ListTemplate,
-            FormTemplate,
-            DashboardTemplate,
-        )
-        
-        # Chat template
-        chat = ChatTemplate(surface_id="chat")
-        chat.add_user_message("Hello!")
-        chat.add_agent_message("Hi there!")
-        chat_messages = chat.to_messages()
-        assert len(chat_messages) >= 2
-        
-        # List template
-        list_template = ListTemplate(surface_id="list", title="Items")
-        list_template.add_item(title="Item 1", description="First")
-        list_template.add_item(title="Item 2", description="Second")
-        list_messages = list_template.to_messages()
-        assert len(list_messages) >= 2
-        
-        # Form template
-        form = FormTemplate(surface_id="form", title="Contact")
-        form.add_text_field(id="name", label="Name")
-        form.add_email_field(id="email", label="Email")
-        form.set_submit_action("submit", "Send")
-        form_messages = form.to_messages()
-        assert len(form_messages) >= 2
-        
-        # Dashboard template
-        dashboard = DashboardTemplate(surface_id="dashboard", title="Dashboard")
-        dashboard.add_panel(id="stats", title="Stats", content="100 items")
-        dashboard.add_panel(id="chart", title="Chart", content="[Chart]")
-        dashboard_messages = dashboard.to_messages()
-        assert len(dashboard_messages) >= 2
-
-    def test_a2a_integration(self):
-        """Test A2A protocol integration."""
-        from praisonaiagents.ui.a2ui import (
-            Surface,
-            create_a2ui_part,
-            is_a2ui_part,
-            A2UI_MIME_TYPE,
-        )
-        
-        # Create surface
-        surface = Surface(surface_id="a2a-test")
-        surface.text(id="msg", text="Hello from A2A")
-        surface.column(id="root", children=["msg"])
-        
-        # Get messages
-        messages = surface.to_messages()
-        
-        # Wrap in A2A DataPart
-        part = create_a2ui_part({"messages": messages})
-        
-        # Verify A2UI part
-        assert is_a2ui_part(part)
-        assert part.metadata["mimeType"] == A2UI_MIME_TYPE
-        assert "messages" in part.data
-        assert len(part.data["messages"]) >= 2
-
-    def test_data_binding_workflow(self):
-        """Test data binding with path references."""
-        from praisonaiagents.ui.a2ui import (
-            Surface,
-            PathBinding,
-            ChildrenTemplate,
-        )
-        
-        surface = Surface(surface_id="data-binding-test")
-        
-        # Add components with data binding
-        surface.text(id="title", text=PathBinding(path="/title"), usage_hint="h1")
-        surface.text(id="item-template", text=PathBinding(path="/name"))
-        
-        # Add list with template
-        surface.list(
-            id="items",
-            children=ChildrenTemplate(component_id="item-template", path="/items"),
-            direction="vertical"
-        )
-        
-        surface.column(id="root", children=["title", "items"])
-        
-        # Set data
-        surface.set_data("title", "My List")
-        surface.set_data("items", [
-            {"name": "Item 1"},
-            {"name": "Item 2"},
-            {"name": "Item 3"},
-        ])
-        
-        messages = surface.to_messages()
-        
-        # Verify data model
-        data_msg = next((m for m in messages if "updateDataModel" in m), None)
-        assert data_msg is not None
-        assert data_msg["updateDataModel"]["value"]["title"] == "My List"
-        assert len(data_msg["updateDataModel"]["value"]["items"]) == 3
-
-
-@pytest.mark.skipif(
-    not os.environ.get("OPENAI_API_KEY"),
-    reason="OPENAI_API_KEY not set"
-)
-class TestA2UIWithLLM:
-    """Tests that require LLM API calls."""
-
-    def test_agent_chat_and_render(self):
-        """Test agent chat followed by UI rendering."""
-        from praisonaiagents import Agent
-        from praisonaiagents.ui.a2ui import A2UIAgent
-        
-        agent = Agent(
-            name="UIAgent",
-            role="UI Generator",
-            goal="Generate helpful responses",
-            llm="gpt-4o-mini"
-        )
-        
-        a2ui_agent = A2UIAgent(agent=agent)
-        
-        # Chat with agent
-        response = a2ui_agent.chat("What is 2+2?")
-        
-        # Render response as UI
-        messages = a2ui_agent.render_text(response, title="Answer")
-        
-        assert len(messages) >= 2
-        assert "createSurface" in messages[0]

src/praisonai-agents/tests/a2ui/test_surface.py

@@ -1,228 +0,0 @@
-"""
-Tests for A2UI Surface Builder
-
-TDD: Write tests first, then implement the Surface class.
-"""
-
-
-class TestSurface:
-    """Test Surface builder class."""
-
-    def test_surface_creation(self):
-        """Test Surface creation with ID and catalog."""
-        from praisonaiagents.ui.a2ui.surface import Surface
-        
-        surface = Surface(surface_id="main")
-        
-        assert surface.surface_id == "main"
-        assert surface.catalog_id is not None  # Should have default
-
-    def test_surface_with_custom_catalog(self):
-        """Test Surface with custom catalog ID."""
-        from praisonaiagents.ui.a2ui.surface import Surface
-        
-        surface = Surface(
-            surface_id="main",
-            catalog_id="https://custom.catalog/v1"
-        )
-        
-        assert surface.catalog_id == "https://custom.catalog/v1"
-
-    def test_add_component(self):
-        """Test adding a component to the surface."""
-        from praisonaiagents.ui.a2ui.surface import Surface
-        from praisonaiagents.ui.a2ui.types import TextComponent
-        
-        surface = Surface(surface_id="main")
-        text = TextComponent(id="title", text="Hello World", usage_hint="h1")
-        
-        surface.add(text)
-        
-        assert len(surface.components) == 1
-        assert surface.components[0].id == "title"
-
-    def test_add_multiple_components(self):
-        """Test adding multiple components."""
-        from praisonaiagents.ui.a2ui.surface import Surface
-        from praisonaiagents.ui.a2ui.types import TextComponent, ButtonComponent
-        
-        surface = Surface(surface_id="main")
-        surface.add(TextComponent(id="title", text="Hello"))
-        surface.add(ButtonComponent(id="btn", child="btn-text"))
-        
-        assert len(surface.components) == 2
-
-    def test_set_data(self):
-        """Test setting data model values."""
-        from praisonaiagents.ui.a2ui.surface import Surface
-        
-        surface = Surface(surface_id="main")
-        surface.set_data("title", "Hello World")
-        surface.set_data("count", 42)
-        
-        assert surface.data["title"] == "Hello World"
-        assert surface.data["count"] == 42
-
-    def test_set_nested_data(self):
-        """Test setting nested data model values."""
-        from praisonaiagents.ui.a2ui.surface import Surface
-        
-        surface = Surface(surface_id="main")
-        surface.set_data("user", {"name": "John", "age": 30})
-        
-        assert surface.data["user"]["name"] == "John"
-
-    def test_to_messages_creates_surface(self):
-        """Test to_messages generates CreateSurface message."""
-        from praisonaiagents.ui.a2ui.surface import Surface
-        from praisonaiagents.ui.a2ui.types import TextComponent
-        
-        surface = Surface(surface_id="main")
-        surface.add(TextComponent(id="title", text="Hello"))
-        
-        messages = surface.to_messages()
-        
-        assert len(messages) >= 1
-        assert "createSurface" in messages[0]
-        assert messages[0]["createSurface"]["surfaceId"] == "main"
-
-    def test_to_messages_includes_components(self):
-        """Test to_messages includes UpdateComponents message."""
-        from praisonaiagents.ui.a2ui.surface import Surface
-        from praisonaiagents.ui.a2ui.types import TextComponent
-        
-        surface = Surface(surface_id="main")
-        surface.add(TextComponent(id="title", text="Hello"))
-        
-        messages = surface.to_messages()
-        
-        # Should have createSurface and updateComponents
-        assert len(messages) >= 2
-        update_msg = next((m for m in messages if "updateComponents" in m), None)
-        assert update_msg is not None
-        assert len(update_msg["updateComponents"]["components"]) == 1
-
-    def test_to_messages_includes_data_model(self):
-        """Test to_messages includes UpdateDataModel message when data is set."""
-        from praisonaiagents.ui.a2ui.surface import Surface
-        from praisonaiagents.ui.a2ui.types import TextComponent
-        
-        surface = Surface(surface_id="main")
-        surface.add(TextComponent(id="title", text="Hello"))
-        surface.set_data("title", "Hello World")
-        
-        messages = surface.to_messages()
-        
-        # Should have createSurface, updateComponents, and updateDataModel
-        assert len(messages) == 3
-        data_msg = next((m for m in messages if "updateDataModel" in m), None)
-        assert data_msg is not None
-        assert data_msg["updateDataModel"]["value"]["title"] == "Hello World"
-
-    def test_to_json(self):
-        """Test to_json returns JSON string."""
-        from praisonaiagents.ui.a2ui.surface import Surface
-        from praisonaiagents.ui.a2ui.types import TextComponent
-        import json
-        
-        surface = Surface(surface_id="main")
-        surface.add(TextComponent(id="title", text="Hello"))
-        
-        json_str = surface.to_json()
-        
-        # Should be valid JSON
-        parsed = json.loads(json_str)
-        assert isinstance(parsed, list)
-        assert len(parsed) >= 2
-
-
-class TestSurfaceChaining:
-    """Test Surface method chaining."""
-
-    def test_add_returns_self(self):
-        """Test add() returns self for chaining."""
-        from praisonaiagents.ui.a2ui.surface import Surface
-        from praisonaiagents.ui.a2ui.types import TextComponent
-        
-        surface = Surface(surface_id="main")
-        result = surface.add(TextComponent(id="title", text="Hello"))
-        
-        assert result is surface
-
-    def test_set_data_returns_self(self):
-        """Test set_data() returns self for chaining."""
-        from praisonaiagents.ui.a2ui.surface import Surface
-        
-        surface = Surface(surface_id="main")
-        result = surface.set_data("key", "value")
-        
-        assert result is surface
-
-    def test_fluent_api(self):
-        """Test fluent API for building surfaces."""
-        from praisonaiagents.ui.a2ui.surface import Surface
-        from praisonaiagents.ui.a2ui.types import TextComponent, ButtonComponent
-        
-        surface = (
-            Surface(surface_id="main")
-            .add(TextComponent(id="title", text="Hello"))
-            .add(ButtonComponent(id="btn", child="btn-text"))
-            .set_data("title", "Hello World")
-        )
-        
-        assert len(surface.components) == 2
-        assert surface.data["title"] == "Hello World"
-
-
-class TestSurfaceHelpers:
-    """Test Surface helper methods."""
-
-    def test_text_helper(self):
-        """Test text() helper method."""
-        from praisonaiagents.ui.a2ui.surface import Surface
-        
-        surface = Surface(surface_id="main")
-        surface.text(id="title", text="Hello", usage_hint="h1")
-        
-        assert len(surface.components) == 1
-        assert surface.components[0].id == "title"
-
-    def test_button_helper(self):
-        """Test button() helper method."""
-        from praisonaiagents.ui.a2ui.surface import Surface
-        
-        surface = Surface(surface_id="main")
-        surface.button(id="submit", child="submit-text", action_name="submit")
-        
-        assert len(surface.components) == 1
-        assert surface.components[0].id == "submit"
-
-    def test_card_helper(self):
-        """Test card() helper method."""
-        from praisonaiagents.ui.a2ui.surface import Surface
-        
-        surface = Surface(surface_id="main")
-        surface.card(id="card-1", child="card-content")
-        
-        assert len(surface.components) == 1
-        assert surface.components[0].id == "card-1"
-
-    def test_row_helper(self):
-        """Test row() helper method."""
-        from praisonaiagents.ui.a2ui.surface import Surface
-        
-        surface = Surface(surface_id="main")
-        surface.row(id="row-1", children=["col-1", "col-2"])
-        
-        assert len(surface.components) == 1
-        assert surface.components[0].id == "row-1"
-
-    def test_column_helper(self):
-        """Test column() helper method."""
-        from praisonaiagents.ui.a2ui.surface import Surface
-        
-        surface = Surface(surface_id="main")
-        surface.column(id="col-1", children=["item-1", "item-2"])
-        
-        assert len(surface.components) == 1
-        assert surface.components[0].id == "col-1"

src/praisonai-agents/tests/a2ui/test_templates.py

@@ -1,210 +0,0 @@
-"""
-Tests for A2UI Templates
-
-TDD: Write tests first, then implement the templates.
-"""
-
-
-class TestChatTemplate:
-    """Test ChatTemplate for conversational UIs."""
-
-    def test_chat_template_creation(self):
-        """Test ChatTemplate creation."""
-        from praisonaiagents.ui.a2ui.templates import ChatTemplate
-        
-        template = ChatTemplate(surface_id="chat")
-        
-        assert template.surface_id == "chat"
-
-    def test_add_user_message(self):
-        """Test adding a user message."""
-        from praisonaiagents.ui.a2ui.templates import ChatTemplate
-        
-        template = ChatTemplate(surface_id="chat")
-        template.add_user_message("Hello!")
-        
-        assert len(template.messages) == 1
-        assert template.messages[0]["role"] == "user"
-        assert template.messages[0]["content"] == "Hello!"
-
-    def test_add_agent_message(self):
-        """Test adding an agent message."""
-        from praisonaiagents.ui.a2ui.templates import ChatTemplate
-        
-        template = ChatTemplate(surface_id="chat")
-        template.add_agent_message("Hi there!")
-        
-        assert len(template.messages) == 1
-        assert template.messages[0]["role"] == "agent"
-
-    def test_to_messages(self):
-        """Test generating A2UI messages."""
-        from praisonaiagents.ui.a2ui.templates import ChatTemplate
-        
-        template = ChatTemplate(surface_id="chat")
-        template.add_user_message("Hello!")
-        template.add_agent_message("Hi there!")
-        
-        messages = template.to_messages()
-        
-        assert isinstance(messages, list)
-        assert len(messages) >= 2  # createSurface + updateComponents
-
-
-class TestListTemplate:
-    """Test ListTemplate for displaying lists of items."""
-
-    def test_list_template_creation(self):
-        """Test ListTemplate creation."""
-        from praisonaiagents.ui.a2ui.templates import ListTemplate
-        
-        template = ListTemplate(
-            surface_id="list",
-            title="My Items"
-        )
-        
-        assert template.surface_id == "list"
-        assert template.title == "My Items"
-
-    def test_add_item(self):
-        """Test adding an item to the list."""
-        from praisonaiagents.ui.a2ui.templates import ListTemplate
-        
-        template = ListTemplate(surface_id="list", title="Items")
-        template.add_item(title="Item 1", description="First item")
-        
-        assert len(template.items) == 1
-        assert template.items[0]["title"] == "Item 1"
-
-    def test_add_item_with_image(self):
-        """Test adding an item with an image."""
-        from praisonaiagents.ui.a2ui.templates import ListTemplate
-        
-        template = ListTemplate(surface_id="list", title="Items")
-        template.add_item(
-            title="Item 1",
-            description="First item",
-            image_url="https://example.com/image.jpg"
-        )
-        
-        assert template.items[0]["image_url"] == "https://example.com/image.jpg"
-
-    def test_to_messages(self):
-        """Test generating A2UI messages."""
-        from praisonaiagents.ui.a2ui.templates import ListTemplate
-        
-        template = ListTemplate(surface_id="list", title="Items")
-        template.add_item(title="Item 1", description="First")
-        template.add_item(title="Item 2", description="Second")
-        
-        messages = template.to_messages()
-        
-        assert isinstance(messages, list)
-        assert len(messages) >= 2
-
-
-class TestFormTemplate:
-    """Test FormTemplate for data collection forms."""
-
-    def test_form_template_creation(self):
-        """Test FormTemplate creation."""
-        from praisonaiagents.ui.a2ui.templates import FormTemplate
-        
-        template = FormTemplate(
-            surface_id="form",
-            title="Contact Form"
-        )
-        
-        assert template.surface_id == "form"
-        assert template.title == "Contact Form"
-
-    def test_add_text_field(self):
-        """Test adding a text field."""
-        from praisonaiagents.ui.a2ui.templates import FormTemplate
-        
-        template = FormTemplate(surface_id="form", title="Form")
-        template.add_text_field(id="name", label="Name")
-        
-        assert len(template.fields) == 1
-        assert template.fields[0]["id"] == "name"
-        assert template.fields[0]["type"] == "text"
-
-    def test_add_number_field(self):
-        """Test adding a number field."""
-        from praisonaiagents.ui.a2ui.templates import FormTemplate
-        
-        template = FormTemplate(surface_id="form", title="Form")
-        template.add_number_field(id="age", label="Age")
-        
-        assert template.fields[0]["type"] == "number"
-
-    def test_add_email_field(self):
-        """Test adding an email field."""
-        from praisonaiagents.ui.a2ui.templates import FormTemplate
-        
-        template = FormTemplate(surface_id="form", title="Form")
-        template.add_email_field(id="email", label="Email")
-        
-        assert template.fields[0]["type"] == "email"
-
-    def test_set_submit_action(self):
-        """Test setting submit action."""
-        from praisonaiagents.ui.a2ui.templates import FormTemplate
-        
-        template = FormTemplate(surface_id="form", title="Form")
-        template.set_submit_action("submit_form", "Send")
-        
-        assert template.submit_action == "submit_form"
-        assert template.submit_label == "Send"
-
-    def test_to_messages(self):
-        """Test generating A2UI messages."""
-        from praisonaiagents.ui.a2ui.templates import FormTemplate
-        
-        template = FormTemplate(surface_id="form", title="Form")
-        template.add_text_field(id="name", label="Name")
-        template.add_email_field(id="email", label="Email")
-        
-        messages = template.to_messages()
-        
-        assert isinstance(messages, list)
-        assert len(messages) >= 2
-
-
-class TestDashboardTemplate:
-    """Test DashboardTemplate for multi-panel dashboards."""
-
-    def test_dashboard_template_creation(self):
-        """Test DashboardTemplate creation."""
-        from praisonaiagents.ui.a2ui.templates import DashboardTemplate
-        
-        template = DashboardTemplate(
-            surface_id="dashboard",
-            title="My Dashboard"
-        )
-        
-        assert template.surface_id == "dashboard"
-        assert template.title == "My Dashboard"
-
-    def test_add_panel(self):
-        """Test adding a panel."""
-        from praisonaiagents.ui.a2ui.templates import DashboardTemplate
-        
-        template = DashboardTemplate(surface_id="dashboard", title="Dashboard")
-        template.add_panel(id="panel-1", title="Panel 1", content="Some content")
-        
-        assert len(template.panels) == 1
-        assert template.panels[0]["id"] == "panel-1"
-
-    def test_to_messages(self):
-        """Test generating A2UI messages."""
-        from praisonaiagents.ui.a2ui.templates import DashboardTemplate
-        
-        template = DashboardTemplate(surface_id="dashboard", title="Dashboard")
-        template.add_panel(id="panel-1", title="Panel 1", content="Content 1")
-        template.add_panel(id="panel-2", title="Panel 2", content="Content 2")
-        
-        messages = template.to_messages()
-        
-        assert isinstance(messages, list)
-        assert len(messages) >= 2

src/praisonai-agents/tests/a2ui/test_types.py

@@ -1,318 +0,0 @@
-"""
-Tests for A2UI Types Module
-
-TDD: Write tests first, then implement the types.
-"""
-
-
-
-class TestA2UIMessageTypes:
-    """Test A2UI message types: CreateSurface, UpdateComponents, UpdateDataModel, DeleteSurface."""
-
-    def test_create_surface_message(self):
-        """Test CreateSurface message creation."""
-        from praisonaiagents.ui.a2ui.types import CreateSurfaceMessage
-        
-        msg = CreateSurfaceMessage(
-            surface_id="main",
-            catalog_id="https://a2ui.dev/standard"
-        )
-        
-        assert msg.surface_id == "main"
-        assert msg.catalog_id == "https://a2ui.dev/standard"
-        
-        # Test serialization
-        data = msg.to_dict()
-        assert "createSurface" in data
-        assert data["createSurface"]["surfaceId"] == "main"
-        assert data["createSurface"]["catalogId"] == "https://a2ui.dev/standard"
-
-    def test_update_components_message(self):
-        """Test UpdateComponents message creation."""
-        from praisonaiagents.ui.a2ui.types import UpdateComponentsMessage, TextComponent
-        
-        text = TextComponent(id="title", text="Hello World", usage_hint="h1")
-        msg = UpdateComponentsMessage(
-            surface_id="main",
-            components=[text]
-        )
-        
-        assert msg.surface_id == "main"
-        assert len(msg.components) == 1
-        
-        # Test serialization
-        data = msg.to_dict()
-        assert "updateComponents" in data
-        assert data["updateComponents"]["surfaceId"] == "main"
-        assert len(data["updateComponents"]["components"]) == 1
-
-    def test_update_data_model_message(self):
-        """Test UpdateDataModel message creation."""
-        from praisonaiagents.ui.a2ui.types import UpdateDataModelMessage
-        
-        msg = UpdateDataModelMessage(
-            surface_id="main",
-            path="/",
-            op="replace",
-            value={"title": "Hello", "items": []}
-        )
-        
-        assert msg.surface_id == "main"
-        assert msg.path == "/"
-        assert msg.op == "replace"
-        assert msg.value == {"title": "Hello", "items": []}
-        
-        # Test serialization
-        data = msg.to_dict()
-        assert "updateDataModel" in data
-        assert data["updateDataModel"]["surfaceId"] == "main"
-
-    def test_delete_surface_message(self):
-        """Test DeleteSurface message creation."""
-        from praisonaiagents.ui.a2ui.types import DeleteSurfaceMessage
-        
-        msg = DeleteSurfaceMessage(surface_id="main")
-        
-        assert msg.surface_id == "main"
-        
-        # Test serialization
-        data = msg.to_dict()
-        assert "deleteSurface" in data
-        assert data["deleteSurface"]["surfaceId"] == "main"
-
-
-class TestA2UIComponentTypes:
-    """Test A2UI component types."""
-
-    def test_text_component(self):
-        """Test Text component creation."""
-        from praisonaiagents.ui.a2ui.types import TextComponent
-        
-        text = TextComponent(id="title", text="Hello World", usage_hint="h1")
-        
-        assert text.id == "title"
-        assert text.text == "Hello World"
-        assert text.usage_hint == "h1"
-        
-        # Test serialization
-        data = text.to_dict()
-        assert data["id"] == "title"
-        assert data["component"] == "Text"
-        assert data["text"] == "Hello World"
-        assert data["usageHint"] == "h1"
-
-    def test_text_component_with_path(self):
-        """Test Text component with data binding path."""
-        from praisonaiagents.ui.a2ui.types import TextComponent, PathBinding
-        
-        text = TextComponent(id="title", text=PathBinding(path="/title"))
-        
-        data = text.to_dict()
-        assert data["text"] == {"path": "/title"}
-
-    def test_image_component(self):
-        """Test Image component creation."""
-        from praisonaiagents.ui.a2ui.types import ImageComponent
-        
-        img = ImageComponent(
-            id="hero",
-            url="https://example.com/image.jpg",
-            fit="cover",
-            usage_hint="header"
-        )
-        
-        assert img.id == "hero"
-        assert img.url == "https://example.com/image.jpg"
-        assert img.fit == "cover"
-        
-        data = img.to_dict()
-        assert data["component"] == "Image"
-        assert data["url"] == "https://example.com/image.jpg"
-
-    def test_button_component(self):
-        """Test Button component creation."""
-        from praisonaiagents.ui.a2ui.types import ButtonComponent, Action, ActionContext
-        
-        action = Action(
-            name="submit",
-            context=[
-                ActionContext(key="value", value="test")
-            ]
-        )
-        btn = ButtonComponent(
-            id="submit-btn",
-            child="submit-text",
-            primary=True,
-            action=action
-        )
-        
-        assert btn.id == "submit-btn"
-        assert btn.child == "submit-text"
-        assert btn.primary is True
-        
-        data = btn.to_dict()
-        assert data["component"] == "Button"
-        assert data["action"]["name"] == "submit"
-
-    def test_card_component(self):
-        """Test Card component creation."""
-        from praisonaiagents.ui.a2ui.types import CardComponent
-        
-        card = CardComponent(id="card-1", child="card-content")
-        
-        assert card.id == "card-1"
-        assert card.child == "card-content"
-        
-        data = card.to_dict()
-        assert data["component"] == "Card"
-        assert data["child"] == "card-content"
-
-    def test_row_component(self):
-        """Test Row component creation."""
-        from praisonaiagents.ui.a2ui.types import RowComponent
-        
-        row = RowComponent(id="row-1", children=["col-1", "col-2"])
-        
-        assert row.id == "row-1"
-        assert row.children == ["col-1", "col-2"]
-        
-        data = row.to_dict()
-        assert data["component"] == "Row"
-        assert data["children"] == ["col-1", "col-2"]
-
-    def test_column_component(self):
-        """Test Column component creation."""
-        from praisonaiagents.ui.a2ui.types import ColumnComponent
-        
-        col = ColumnComponent(id="col-1", children=["item-1", "item-2"])
-        
-        assert col.id == "col-1"
-        assert col.children == ["item-1", "item-2"]
-        
-        data = col.to_dict()
-        assert data["component"] == "Column"
-
-    def test_list_component(self):
-        """Test List component creation."""
-        from praisonaiagents.ui.a2ui.types import ListComponent, ChildrenTemplate
-        
-        template = ChildrenTemplate(component_id="item-template", path="/items")
-        lst = ListComponent(id="item-list", direction="vertical", children=template)
-        
-        assert lst.id == "item-list"
-        assert lst.direction == "vertical"
-        
-        data = lst.to_dict()
-        assert data["component"] == "List"
-        assert data["children"]["componentId"] == "item-template"
-        assert data["children"]["path"] == "/items"
-
-    def test_text_field_component(self):
-        """Test TextField component creation."""
-        from praisonaiagents.ui.a2ui.types import TextFieldComponent, PathBinding
-        
-        field = TextFieldComponent(
-            id="name-field",
-            label="Name",
-            text=PathBinding(path="/name"),
-            field_type="text"
-        )
-        
-        assert field.id == "name-field"
-        assert field.label == "Name"
-        
-        data = field.to_dict()
-        assert data["component"] == "TextField"
-        assert data["label"] == "Name"
-        assert data["text"] == {"path": "/name"}
-
-    def test_checkbox_component(self):
-        """Test CheckBox component creation."""
-        from praisonaiagents.ui.a2ui.types import CheckBoxComponent, PathBinding
-        
-        cb = CheckBoxComponent(
-            id="agree-cb",
-            label="I agree",
-            checked=PathBinding(path="/agreed")
-        )
-        
-        data = cb.to_dict()
-        assert data["component"] == "CheckBox"
-        assert data["label"] == "I agree"
-
-    def test_slider_component(self):
-        """Test Slider component creation."""
-        from praisonaiagents.ui.a2ui.types import SliderComponent
-        
-        slider = SliderComponent(
-            id="volume",
-            min_value=0,
-            max_value=100,
-            value=50
-        )
-        
-        data = slider.to_dict()
-        assert data["component"] == "Slider"
-        assert data["min"] == 0
-        assert data["max"] == 100
-
-
-class TestDataBindingTypes:
-    """Test data binding types."""
-
-    def test_path_binding(self):
-        """Test PathBinding for data model references."""
-        from praisonaiagents.ui.a2ui.types import PathBinding
-        
-        path = PathBinding(path="/user/name")
-        
-        assert path.path == "/user/name"
-        assert path.to_dict() == {"path": "/user/name"}
-
-    def test_string_or_path_literal(self):
-        """Test StringOrPath with literal string."""
-        from praisonaiagents.ui.a2ui.types import resolve_string_or_path
-        
-        result = resolve_string_or_path("Hello")
-        assert result == "Hello"
-
-    def test_string_or_path_binding(self):
-        """Test StringOrPath with path binding."""
-        from praisonaiagents.ui.a2ui.types import resolve_string_or_path, PathBinding
-        
-        result = resolve_string_or_path(PathBinding(path="/title"))
-        assert result == {"path": "/title"}
-
-
-class TestActionTypes:
-    """Test action types for button events."""
-
-    def test_action_creation(self):
-        """Test Action creation."""
-        from praisonaiagents.ui.a2ui.types import Action, ActionContext
-        
-        action = Action(
-            name="book_restaurant",
-            context=[
-                ActionContext(key="restaurantName", value="The Fancy Place"),
-                ActionContext(key="address", value="123 Main St")
-            ]
-        )
-        
-        assert action.name == "book_restaurant"
-        assert len(action.context) == 2
-        
-        data = action.to_dict()
-        assert data["name"] == "book_restaurant"
-        assert len(data["context"]) == 2
-        assert data["context"][0]["key"] == "restaurantName"
-
-    def test_action_context_with_path(self):
-        """Test ActionContext with path binding."""
-        from praisonaiagents.ui.a2ui.types import ActionContext, PathBinding
-        
-        ctx = ActionContext(key="name", value=PathBinding(path="/restaurant/name"))
-        
-        data = ctx.to_dict()
-        assert data["key"] == "name"
-        assert data["value"] == {"path": "/restaurant/name"}

src/praisonai-agents/tests/agui/test_streaming.py

@@ -98,6 +98,38 @@ class TestToolCallStreaming:
         assert event.content == "Search results..."
 
 
+class TestStreamEventBridge:
+    """Test PraisonAI stream_emitter → AG-UI event conversion."""
+
+    def test_delta_text_to_agui(self):
+        from praisonaiagents.streaming.events import StreamEvent, StreamEventType
+        from praisonaiagents.ui.agui.streaming import EventBuffer, stream_event_to_agui_events
+
+        buffer = EventBuffer()
+        event = StreamEvent(type=StreamEventType.DELTA_TEXT, content="Hello")
+        agui_events = stream_event_to_agui_events(event, "msg-1", buffer)
+
+        assert len(agui_events) == 2
+        assert agui_events[0].type == "TEXT_MESSAGE_START"
+        assert agui_events[1].type == "TEXT_MESSAGE_CONTENT"
+        assert agui_events[1].delta == "Hello"
+
+    def test_tool_call_start_to_agui(self):
+        from praisonaiagents.streaming.events import StreamEvent, StreamEventType
+        from praisonaiagents.ui.agui.streaming import EventBuffer, stream_event_to_agui_events
+
+        buffer = EventBuffer()
+        event = StreamEvent(
+            type=StreamEventType.TOOL_CALL_START,
+            tool_call={"id": "tc-1", "name": "search", "arguments": {"q": "test"}},
+        )
+        agui_events = stream_event_to_agui_events(event, "msg-1", buffer)
+
+        assert len(agui_events) == 3
+        assert agui_events[0].type == "TOOL_CALL_START"
+        assert agui_events[0].tool_call_name == "search"
+
+
 class TestRunLifecycleEvents:
     """Test run lifecycle events."""
     

src/praisonai-agents/tests/integration/test_allowed_tools_filter_integration.py

@@ -0,0 +1,328 @@
+"""Integration tests for the ALLOWED_TOOLS filter module."""
+import os
+import pytest
+from unittest.mock import patch, MagicMock
+from praisonai.cli.features.agents import MultiAgentHandler
+
+
+class TestAllowedToolsFilterRegistryIntegration:
+    """Integration tests for ALLOWED_TOOLS filter with the tool registry."""
+    
+    def test_registry_allowed_filter_unset(self):
+        """Test registry filter behavior when ALLOWED_TOOLS is unset."""
+        from praisonaiagents.tools.registry import get_registry
+        
+        with patch.dict(os.environ, {}, clear=True):
+            registry = get_registry()
+            
+            # Mock some tools in the registry
+            with patch.object(registry, 'list_tools', return_value=['search_web', 'send_email']):
+                filtered_tools = registry.list_tools_with_allowed_filter()
+                
+                # Should return all tools when unset
+                assert set(filtered_tools) == {'search_web', 'send_email'}
+    
+    def test_registry_allowed_filter_with_whitelist(self):
+        """Test registry filter with ALLOWED_TOOLS whitelist."""
+        from praisonaiagents.tools.registry import get_registry
+        
+        with patch.dict(os.environ, {
+            "ALLOWED_TOOLS": "search_web,send_email,unknown_tool",
+            "CI": "false",
+        }):
+            registry = get_registry()
+            
+            # Mock some tools in the registry
+            with patch.object(registry, 'list_tools', return_value=['search_web', 'send_email', 'extract_pdf']):
+                filtered_tools = registry.list_tools_with_allowed_filter()
+                
+                # Should only include whitelisted tools that are available
+                assert set(filtered_tools) == {'search_web', 'send_email'}
+    
+    def test_registry_allowed_filter_empty_result(self):
+        """Test registry filter when no tools match whitelist."""
+        from praisonaiagents.tools.registry import get_registry
+        
+        with patch.dict(os.environ, {
+            "ALLOWED_TOOLS": "nonexistent_tool1,nonexistent_tool2",
+            "CI": "false",
+        }):
+            registry = get_registry()
+            
+            # Mock some tools in the registry
+            with patch.object(registry, 'list_tools', return_value=['search_web', 'send_email']):
+                filtered_tools = registry.list_tools_with_allowed_filter()
+                
+                # Should return empty list when no matches
+                assert filtered_tools == []
+    
+    def test_registry_allowed_filter_import_error_fallback(self):
+        """Test registry filter fallback when import fails."""
+        from praisonaiagents.tools.registry import get_registry
+        
+        registry = get_registry()
+        
+        with patch.object(registry, 'list_tools', return_value=['search_web', 'send_email']):
+            # Mock import error for the filter
+            with patch('praisonaiagents.tools.registry.filter_tools_with_allowed_tools', side_effect=ImportError("Filter not available")):
+                filtered_tools = registry.list_tools_with_allowed_filter()
+                
+                # Should fallback to all tools
+                assert set(filtered_tools) == {'search_web', 'send_email'}
+    
+    def test_registry_allowed_filter_value_error_propagation(self):
+        """Test that ValueError from filter is properly propagated."""
+        from praisonaiagents.tools.registry import get_registry
+        
+        with patch.dict(os.environ, {"ALLOWED_TOOLS": ""}):  # Empty string should cause ValueError
+            registry = get_registry()
+            
+            with patch.object(registry, 'list_tools', return_value=['search_web', 'send_email']):
+                with pytest.raises(ValueError, match="cannot be empty"):
+                    registry.list_tools_with_allowed_filter()
+    
+    def test_registry_allowed_filter_general_error_fallback(self):
+        """Test registry filter fallback on general errors."""
+        from praisonaiagents.tools.registry import get_registry
+        
+        registry = get_registry()
+        
+        with patch.object(registry, 'list_tools', return_value=['search_web', 'send_email']):
+            # Mock a generic error during filtering
+            with patch('praisonaiagents.tools.registry.filter_tools_with_allowed_tools', side_effect=RuntimeError("Some error")):
+                filtered_tools = registry.list_tools_with_allowed_filter()
+                
+                # Should fallback to all tools on general errors
+                assert set(filtered_tools) == {'search_web', 'send_email'}
+
+
+class TestAllowedToolsFilterGlobalRegistryPath:
+    """Test ALLOWED_TOOLS filter through the global registry function."""
+    
+    def test_global_list_tools_with_allowed_filter(self):
+        """Test the global list_tools_with_allowed_filter function."""
+        from praisonaiagents.tools import list_tools_with_allowed_filter
+        from praisonaiagents.tools.registry import get_registry
+        
+        with patch.dict(os.environ, {"ALLOWED_TOOLS": "search_web,send_email"}):
+            registry = get_registry()
+            
+            with patch.object(registry, 'list_tools_with_allowed_filter', return_value=['search_web', 'send_email']) as mock_filter:
+                result = list_tools_with_allowed_filter()
+                
+                # Should call the registry method
+                mock_filter.assert_called_once_with(None)
+                assert result == ['search_web', 'send_email']
+    
+    def test_global_backward_compatibility_alias(self):
+        """Test the backward compatibility alias for hermes filter."""
+        from praisonaiagents.tools import list_tools_with_hermes_filter, list_tools_with_allowed_filter
+        from praisonaiagents.tools.registry import get_registry
+        
+        registry = get_registry()
+        
+        with patch.object(registry, 'list_tools_with_allowed_filter', return_value=['search_web']) as mock_filter:
+            # Both should call the same underlying function
+            result1 = list_tools_with_allowed_filter()
+            result2 = list_tools_with_hermes_filter()
+            
+            assert result1 == result2 == ['search_web']
+            assert mock_filter.call_count == 2
+
+
+class TestAllowedToolsFilterCLIIntegration:
+    """Integration tests for ALLOWED_TOOLS filter with CLI agent handler."""
+    
+    def test_cli_agents_load_tools_no_filter(self):
+        """Test CLI agent tool loading without ALLOWED_TOOLS filter."""
+        handler = MultiAgentHandler(verbose=False)
+        
+        with patch.dict(os.environ, {}, clear=True):
+            # Mock the tools import to avoid actual import dependencies
+            with patch('praisonai.cli.features.agents.internet_search'), \
+                 patch('praisonai.cli.features.agents.read_file'):
+                
+                tools = handler._load_tools(["internet_search", "read_file"])
+                
+                # Should load all requested tools when no filter
+                assert len(tools) == 2
+    
+    def test_cli_agents_load_tools_with_filter(self):
+        """Test CLI agent tool loading with ALLOWED_TOOLS filter."""
+        handler = MultiAgentHandler(verbose=False)
+        
+        with patch.dict(os.environ, {"ALLOWED_TOOLS": "internet_search"}):
+            # Mock the tools and filter
+            mock_internet_search = MagicMock()
+            mock_read_file = MagicMock()
+            
+            with patch('praisonai.cli.features.agents.internet_search', mock_internet_search), \
+                 patch('praisonai.cli.features.agents.read_file', mock_read_file):
+                
+                tools = handler._load_tools(["internet_search", "read_file"])
+                
+                # Should only load tools allowed by filter
+                assert len(tools) == 1
+                assert mock_internet_search in tools
+                assert mock_read_file not in tools
+    
+    def test_cli_agents_load_tools_filter_error_handling(self):
+        """Test CLI tool loading with ALLOWED_TOOLS filter error handling."""
+        handler = MultiAgentHandler(verbose=False)
+        requested = ["internet_search", "read_file"]
+        
+        with patch('os.environ.get', return_value=None):
+            with patch('praisonai.cli.features.agents.internet_search'), \
+                 patch('praisonai.cli.features.agents.read_file'):
+                baseline_tools = handler._load_tools(requested)
+        
+        with patch.dict(os.environ, {"ALLOWED_TOOLS": ""}):
+            with pytest.raises(ValueError):
+                handler._load_tools(requested)
+    
+    def test_cli_agents_load_tools_import_error_fallback(self):
+        """Test CLI tool loading fallback when ALLOWED_TOOLS filter import fails."""
+        handler = MultiAgentHandler(verbose=False)
+        
+        # Mock ImportError for the filter module
+        with patch('praisonai.cli.features.agents.filter_tools_with_allowed_tools', side_effect=ImportError("Filter module not available")):
+            with patch('praisonai.cli.features.agents.internet_search'), \
+                 patch('praisonai.cli.features.agents.read_file'):
+                
+                tools = handler._load_tools(["internet_search", "read_file"])
+                
+                # Should load all tools on import error
+                assert len(tools) == 2
+    
+    def test_cli_agents_load_tools_filter_value_error_propagation(self):
+        """Test that ValueError from filter is properly propagated in CLI."""
+        handler = MultiAgentHandler(verbose=False)
+        
+        with patch.dict(os.environ, {"ALLOWED_TOOLS": ""}):  # Empty string causes ValueError
+            with pytest.raises(ValueError, match="cannot be empty"):
+                handler._load_tools(["internet_search"])
+
+
+class TestAllowedToolsFilterSpecSemantics:
+    """Test ALLOWED_TOOLS filter specification semantics."""
+    
+    def test_spec_semantics_unset_all_tools_visible(self):
+        """Test spec: unset ALLOWED_TOOLS means all tools visible with warning."""
+        from praisonaiagents.allowed_tools_filter import AllowedToolsFilter
+        
+        with patch.dict(os.environ, {}, clear=True):
+            filter_instance = AllowedToolsFilter()
+            available_tools = {"search", "send_message", "extract_pdf"}
+            
+            import logging
+            with patch.object(logging.getLogger(), 'warning') as mock_warn:
+                filtered = filter_instance.filter_tools(available_tools)
+            
+            # Should return all tools
+            assert filtered == available_tools
+            
+            # Should warn about unset whitelist
+            mock_warn.assert_called_once()
+            assert "Tool whitelist is unset" in mock_warn.call_args[0][0]
+    
+    def test_spec_semantics_empty_string_error(self):
+        """Test spec: empty ALLOWED_TOOLS string is an error."""
+        from praisonaiagents.allowed_tools_filter import AllowedToolsFilter
+        
+        with patch.dict(os.environ, {"ALLOWED_TOOLS": ""}):
+            with pytest.raises(ValueError, match="ALLOWED_TOOLS cannot be empty"):
+                AllowedToolsFilter()
+    
+    def test_spec_semantics_whitelist_only_specified_visible(self):
+        """Test spec: with ALLOWED_TOOLS values, only whitelisted tools visible."""
+        from praisonaiagents.allowed_tools_filter import AllowedToolsFilter
+        
+        with patch.dict(os.environ, {"ALLOWED_TOOLS": "search,send_message"}):
+            filter_instance = AllowedToolsFilter()
+            available_tools = {"search", "send_message", "extract_pdf", "analyze_code"}
+            
+            filtered = filter_instance.filter_tools(available_tools)
+            
+            # Should only return whitelisted tools
+            assert filtered == {"search", "send_message"}
+    
+    def test_spec_semantics_unknown_tools_dev_warn_strip(self):
+        """Test spec: unknown tools in dev mode are warned and stripped."""
+        from praisonaiagents.allowed_tools_filter import AllowedToolsFilter
+        
+        with patch.dict(os.environ, {"ALLOWED_TOOLS": "search,unknown_tool", "CI": "false"}):
+            filter_instance = AllowedToolsFilter()
+            available_tools = {"search", "send_message"}
+            
+            import logging
+            with patch.object(logging.getLogger(), 'warning') as mock_warn:
+                filtered = filter_instance.filter_tools(available_tools)
+            
+            # Should return only known tools
+            assert filtered == {"search"}
+            
+            # Should warn about unknown tools
+            mock_warn.assert_called()
+            assert "unknown_tool" in str(mock_warn.call_args[0][2])
+    
+    def test_spec_semantics_unknown_tools_ci_strict_fail(self):
+        """Test spec: unknown tools in CI mode cause strict failure."""
+        from praisonaiagents.allowed_tools_filter import AllowedToolsFilter
+        
+        with patch.dict(os.environ, {"ALLOWED_TOOLS": "search,unknown_tool", "CI": "true"}):
+            filter_instance = AllowedToolsFilter()
+            available_tools = {"search", "send_message"}
+            
+            with pytest.raises(ValueError, match="Unknown tools in ALLOWED_TOOLS"):
+                filter_instance.filter_tools(available_tools)
+    
+    def test_spec_semantics_order_independence(self):
+        """Test spec: whitelist order doesn't matter."""
+        from praisonaiagents.allowed_tools_filter import AllowedToolsFilter
+        
+        available_tools = {"search", "send_message", "extract_pdf"}
+        
+        with patch.dict(os.environ, {"ALLOWED_TOOLS": "search,send_message"}):
+            filter1 = AllowedToolsFilter()
+            result1 = filter1.filter_tools(available_tools)
+        
+        with patch.dict(os.environ, {"ALLOWED_TOOLS": "send_message,search"}):
+            filter2 = AllowedToolsFilter()
+            result2 = filter2.filter_tools(available_tools)
+        
+        # Results should be the same regardless of order
+        assert result1 == result2 == {"search", "send_message"}
+    
+    def test_spec_semantics_duplicate_isolation(self):
+        """Test spec: duplicate tool names in whitelist are handled properly."""
+        from praisonaiagents.allowed_tools_filter import AllowedToolsFilter
+        
+        with patch.dict(os.environ, {"ALLOWED_TOOLS": "search,search,send_message,search"}):
+            filter_instance = AllowedToolsFilter()
+            available_tools = {"search", "send_message", "extract_pdf"}
+            
+            filtered = filter_instance.filter_tools(available_tools)
+            
+            # Duplicates should be de-duplicated
+            assert filtered == {"search", "send_message"}
+            
+            # Whitelist should be a set (no duplicates)
+            whitelist = filter_instance.get_whitelist()
+            assert whitelist == {"search", "send_message"}
+    
+    def test_spec_backward_compatibility_hermes_precedence(self):
+        """Test spec: ALLOWED_TOOLS takes precedence over HERMES_ONLY_TOOLS."""
+        from praisonaiagents.allowed_tools_filter import AllowedToolsFilter
+        
+        with patch.dict(os.environ, {
+            "ALLOWED_TOOLS": "search,send_message",
+            "HERMES_ONLY_TOOLS": "extract_pdf"
+        }):
+            filter_instance = AllowedToolsFilter()
+            available_tools = {"search", "send_message", "extract_pdf"}
+            
+            filtered = filter_instance.filter_tools(available_tools)
+            
+            # Should use ALLOWED_TOOLS, not HERMES_ONLY_TOOLS
+            assert filtered == {"search", "send_message"}
+            assert filter_instance.env_var_name == "ALLOWED_TOOLS"

src/praisonai-agents/tests/unit/agent/test_achat_unified_dispatch.py

@@ -0,0 +1,84 @@
+"""Unit tests for unified achat dispatch (Issue #1703)."""
+
+import asyncio
+from types import SimpleNamespace
+from unittest.mock import AsyncMock, MagicMock, patch
+
+import pytest
+
+from praisonaiagents import Agent
+
+
+def _fake_completion(content: str = "Mock async reply"):
+    return SimpleNamespace(
+        choices=[SimpleNamespace(message=SimpleNamespace(content=content, reasoning_content=None))]
+    )
+
+
+@pytest.mark.asyncio
+async def test_unified_achat_returns_after_single_dispatch():
+    """achat must return after one unified dispatch, not loop forever."""
+    agent = Agent(name="test", instructions="You are helpful", llm="gpt-4o-mini")
+    agent._using_custom_llm = False
+    agent.self_reflect = False
+
+    mock_unified = AsyncMock(return_value=_fake_completion())
+
+    with patch.object(agent, "_execute_unified_achat_completion", mock_unified), patch.object(
+        agent, "_apply_guardrail_with_retry", return_value="Mock async reply"
+    ), patch.object(
+        agent, "_atrigger_after_agent_hook", new_callable=AsyncMock, return_value="Mock async reply"
+    ), patch.object(
+        agent, "_execute_callback_and_display"
+    ), patch.object(
+        agent, "_build_messages", return_value=([{"role": "user", "content": "hi"}], "hi")
+    ), patch.object(
+        agent, "_persist_message"
+    ), patch.object(
+        agent, "_format_tools_for_completion", return_value=[]
+    ):
+        result = await asyncio.wait_for(agent.achat("hi", stream=False), timeout=5)
+
+    assert result == "Mock async reply"
+    mock_unified.assert_called_once()
+
+
+@pytest.mark.asyncio
+async def test_unified_achat_completion_uses_execute_tool_async():
+    """Async unified dispatch must wire execute_tool_async, not sync execute_tool."""
+    agent = Agent(name="test", instructions="You are helpful", llm="gpt-4o-mini")
+    agent._using_custom_llm = False
+
+    mock_dispatcher = MagicMock()
+    mock_dispatcher.achat_completion = AsyncMock(return_value=_fake_completion())
+    agent._unified_dispatcher = mock_dispatcher
+
+    with patch.object(agent, "_apply_guardrail_with_retry", return_value="ok"), patch.object(
+        agent, "_atrigger_after_agent_hook", new_callable=AsyncMock, return_value="ok"
+    ), patch.object(agent, "_execute_callback_and_display"), patch.object(
+        agent, "_build_messages", return_value=([{"role": "user", "content": "hi"}], "hi")
+    ), patch.object(agent, "_persist_message"), patch.object(
+        agent, "_format_tools_for_completion", return_value=[]
+    ):
+        await agent._achat_impl(
+            prompt="hi",
+            temperature=1.0,
+            tools=None,
+            output_json=None,
+            output_pydantic=None,
+            reasoning_steps=False,
+            stream=False,
+            task_name=None,
+            task_description=None,
+            task_id=None,
+            config=None,
+            force_retrieval=False,
+            skip_retrieval=True,
+            attachments=None,
+            _trace_emitter=MagicMock(),
+        )
+
+    call_kwargs = mock_dispatcher.achat_completion.call_args[1]
+    tool_fn = call_kwargs.get("execute_tool_fn")
+    assert tool_fn is not None
+    assert getattr(tool_fn, "__name__", "") == "execute_tool_async"

src/praisonai-agents/tests/unit/agents/test_server_registry_timeout.py

@@ -0,0 +1,128 @@
+"""Test for agent server readiness timeout warning functionality."""
+
+import time
+import threading
+import logging
+import pytest
+from unittest.mock import patch, MagicMock
+from praisonaiagents.agents.agents import _AgentServerRegistry
+
+
+def test_server_readiness_timeout_emits_warning(caplog):
+    """A server that does not start in time must produce a WARNING log entry."""
+    registry = _AgentServerRegistry()
+    
+    # Create a mock FastAPI app that won't signal readiness
+    mock_app = MagicMock()
+    
+    # Register the app but don't set the ready event
+    with patch.dict(registry._apps, {8000: mock_app}):
+        with patch.dict(registry._ready_events, {8000: threading.Event()}):
+            with patch.dict(registry._started, {8000: False}):
+                with patch("threading.Thread.start"):  # Prevent real server spawning
+                    with caplog.at_level(logging.WARNING):
+                        # Use a very short timeout to ensure the test doesn't hang
+                        with patch.dict("os.environ", {"PRAISONAI_SERVER_READY_TIMEOUT": "0.1"}):
+                            result = registry.start_server_if_needed(8000, host="127.0.0.1")
+                    
+                    # The method should still return True but log a warning
+                    assert result is True
+                    assert "8000" in caplog.text
+                    assert "did not become ready" in caplog.text.lower()
+
+
+def test_server_readiness_timeout_configurable_via_env():
+    """The timeout should be configurable via PRAISONAI_SERVER_READY_TIMEOUT."""
+    registry = _AgentServerRegistry()
+    
+    # Create a mock FastAPI app
+    mock_app = MagicMock()
+    
+    # Register the app but don't set the ready event
+    with patch.dict(registry._apps, {8001: mock_app}):
+        with patch.dict(registry._ready_events, {8001: threading.Event()}):
+            with patch.dict(registry._started, {8001: False}):
+                with patch("threading.Thread.start"):  # Prevent real server spawning
+                    # Test with custom timeout
+                    with patch.dict("os.environ", {"PRAISONAI_SERVER_READY_TIMEOUT": "0.05"}):
+                        start_time = time.time()
+                        result = registry.start_server_if_needed(8001, host="127.0.0.1")
+                        duration = time.time() - start_time
+                    
+                    # Should respect the custom timeout and complete quickly
+                    assert result is True
+                    assert duration < 1.0  # Should be much less than default 5s
+
+
+def test_server_readiness_success_no_warning(caplog):
+    """A server that starts in time should not produce any warning."""
+    registry = _AgentServerRegistry()
+    
+    # Create a mock FastAPI app
+    mock_app = MagicMock()
+    ready_event = threading.Event()
+    
+    # Register the app and signal readiness immediately
+    with patch.dict(registry._apps, {8002: mock_app}):
+        with patch.dict(registry._ready_events, {8002: ready_event}):
+            with patch.dict(registry._started, {8002: False}):
+                with patch("threading.Thread.start"):  # Prevent real server spawning
+                    # Set the event immediately to simulate quick startup
+                    ready_event.set()
+                    
+                    with caplog.at_level(logging.WARNING):
+                        result = registry.start_server_if_needed(8002, host="127.0.0.1")
+                    
+                    # Should succeed without warnings
+                    assert result is True
+                    assert "did not become ready" not in caplog.text.lower()
+
+
+def test_default_timeout_value(monkeypatch):
+    """Test that the default timeout is 5.0 seconds when env var is not set."""
+    registry = _AgentServerRegistry()
+    
+    # Create a mock FastAPI app
+    mock_app = MagicMock()
+    ready_event = threading.Event()
+    
+    # Register the app but don't set the ready event
+    with patch.dict(registry._apps, {8003: mock_app}):
+        with patch.dict(registry._ready_events, {8003: ready_event}):
+            with patch.dict(registry._started, {8003: False}):
+                with patch("threading.Thread.start"):  # Prevent real server spawning
+                    # Remove only the specific env var so other env vars (PATH etc.) remain intact
+                    monkeypatch.delenv("PRAISONAI_SERVER_READY_TIMEOUT", raising=False)
+                    with patch("threading.Event.wait") as mock_wait:
+                        mock_wait.return_value = False
+                        result = registry.start_server_if_needed(8003, host="127.0.0.1")
+                    
+                    # Should have called wait with default timeout of 5.0
+                    mock_wait.assert_called_once_with(timeout=5.0)
+                    assert result is True
+
+
+def test_invalid_timeout_env_var_fallback(caplog):
+    """Test that invalid PRAISONAI_SERVER_READY_TIMEOUT falls back to default with warning."""
+    registry = _AgentServerRegistry()
+    
+    # Create a mock FastAPI app
+    mock_app = MagicMock()
+    ready_event = threading.Event()
+    
+    # Register the app but don't set the ready event
+    with patch.dict(registry._apps, {8004: mock_app}):
+        with patch.dict(registry._ready_events, {8004: ready_event}):
+            with patch.dict(registry._started, {8004: False}):
+                with patch("threading.Thread.start"):  # Prevent real server spawning
+                    # Test with invalid timeout value
+                    with patch.dict("os.environ", {"PRAISONAI_SERVER_READY_TIMEOUT": "invalid"}):
+                        with caplog.at_level(logging.WARNING):
+                            with patch("threading.Event.wait") as mock_wait:
+                                mock_wait.return_value = False
+                                result = registry.start_server_if_needed(8004, host="127.0.0.1")
+                    
+                    # Should have fallen back to 5.0 and logged warning
+                    mock_wait.assert_called_once_with(timeout=5.0)
+                    assert result is True
+                    assert "Invalid PRAISONAI_SERVER_READY_TIMEOUT" in caplog.text

src/praisonai-agents/tests/unit/sandbox/__init__.py

@@ -0,0 +1 @@
+# Sandbox tests package

src/praisonai-agents/tests/unit/sandbox/test_sandbox_manager.py

@@ -0,0 +1,199 @@
+"""
+Unit tests for SandboxManager.
+"""
+
+import pytest
+from unittest.mock import Mock, AsyncMock, patch
+from praisonaiagents.sandbox.manager import SandboxManager
+from praisonaiagents.sandbox.config import SandboxConfig
+from praisonaiagents.sandbox.protocols import SandboxResult, SandboxStatus
+
+
+class TestSandboxManager:
+    """Test SandboxManager functionality."""
+
+    def test_init_default_config(self):
+        """Test initialization with default config."""
+        manager = SandboxManager()
+        assert manager.config is not None
+        assert manager.config.sandbox_type == "subprocess"
+
+    def test_init_with_config(self):
+        """Test initialization with custom config."""
+        config = SandboxConfig.docker("python:3.11")
+        manager = SandboxManager(config)
+        assert manager.config == config
+
+    @patch('praisonaiagents.sandbox.manager.SandboxManager._create_sandbox')
+    async def test_context_manager(self, mock_create_sandbox):
+        """Test SandboxManager as async context manager."""
+        mock_sandbox = AsyncMock()
+        mock_create_sandbox.return_value = mock_sandbox
+        
+        manager = SandboxManager()
+        
+        async with manager as sandbox:
+            assert sandbox == mock_sandbox
+            mock_create_sandbox.assert_called_once()
+        
+        # Cleanup should be called
+        mock_sandbox.stop.assert_called_once()
+        mock_sandbox.cleanup.assert_called_once()
+
+    @patch('praisonaiagents.sandbox.manager.SandboxManager._create_sandbox')
+    async def test_context_manager_cleanup_error(self, mock_create_sandbox):
+        """Test context manager cleanup with error."""
+        mock_sandbox = AsyncMock()
+        mock_sandbox.stop.side_effect = Exception("Stop failed")
+        mock_create_sandbox.return_value = mock_sandbox
+        
+        manager = SandboxManager()
+        
+        # Should not raise exception, just log warning
+        async with manager as sandbox:
+            assert sandbox == mock_sandbox
+        
+        mock_sandbox.stop.assert_called_once()
+        mock_sandbox.cleanup.assert_called_once()
+
+    @patch('praisonaiagents.sandbox.manager.SandboxManager._create_sandbox')
+    async def test_run_code_convenience_method(self, mock_create_sandbox):
+        """Test run_code convenience method."""
+        mock_sandbox = AsyncMock()
+        mock_result = SandboxResult(status=SandboxStatus.COMPLETED, stdout="output")
+        mock_sandbox.execute.return_value = mock_result
+        mock_create_sandbox.return_value = mock_sandbox
+        
+        manager = SandboxManager()
+        
+        result = await manager.run_code("print('hello')", language="python")
+        
+        assert result == mock_result
+        mock_sandbox.execute.assert_called_once_with("print('hello')", language="python")
+
+    async def test_create_sandbox_subprocess(self):
+        """Test creating subprocess sandbox."""
+        config = SandboxConfig.subprocess()
+        manager = SandboxManager(config)
+        
+        with patch('praisonai.sandbox.subprocess.SubprocessSandbox') as mock_class:
+            mock_instance = AsyncMock()
+            mock_class.return_value = mock_instance
+            mock_instance.is_available = True
+            
+            sandbox = await manager._create_sandbox()
+            
+            mock_class.assert_called_once_with(config)
+            assert sandbox == mock_instance
+
+    async def test_create_sandbox_docker(self):
+        """Test creating Docker sandbox."""
+        config = SandboxConfig.docker("python:3.11")
+        manager = SandboxManager(config)
+        
+        with patch('praisonai.sandbox.docker.DockerSandbox') as mock_class:
+            mock_instance = AsyncMock()
+            mock_class.return_value = mock_instance
+            mock_instance.is_available = True
+            
+            sandbox = await manager._create_sandbox()
+            
+            mock_class.assert_called_once_with(config)
+            assert sandbox == mock_instance
+
+    async def test_create_sandbox_e2b(self):
+        """Test creating E2B sandbox."""
+        config = SandboxConfig.e2b()
+        manager = SandboxManager(config)
+        
+        with patch('praisonai.sandbox.e2b.E2BSandbox') as mock_class:
+            mock_instance = AsyncMock()
+            mock_class.return_value = mock_instance
+            mock_instance.is_available = True
+            
+            sandbox = await manager._create_sandbox()
+            
+            mock_class.assert_called_once_with(config)
+            assert sandbox == mock_instance
+
+    async def test_create_sandbox_unknown_type(self):
+        """Test creating sandbox with unknown type."""
+        config = SandboxConfig(sandbox_type="unknown")
+        manager = SandboxManager(config)
+        
+        with pytest.raises(ValueError, match="Unknown sandbox type"):
+            await manager._create_sandbox()
+
+    async def test_create_sandbox_unavailable(self):
+        """Test creating sandbox when not available."""
+        config = SandboxConfig.docker("python:3.11")
+        manager = SandboxManager(config)
+        
+        with patch('praisonai.sandbox.docker.DockerSandbox') as mock_class:
+            mock_instance = Mock()
+            mock_instance.is_available = False
+            mock_class.return_value = mock_instance
+            
+            with pytest.raises(RuntimeError, match="not available"):
+                await manager._create_sandbox()
+
+    async def test_create_sandbox_import_error(self):
+        """Test creating sandbox with import error."""
+        config = SandboxConfig.e2b()
+        manager = SandboxManager(config)
+        
+        with patch('importlib.import_module', side_effect=ImportError("Module not found")):
+            with pytest.raises(ImportError, match="Module not found"):
+                await manager._create_sandbox()
+
+    def test_get_available_types(self):
+        """Test getting available sandbox types."""
+        manager = SandboxManager()
+        
+        with patch('praisonaiagents.sandbox.manager.SandboxManager._check_availability') as mock_check:
+            mock_check.side_effect = lambda t: t in ["subprocess", "docker"]
+            
+            available = manager.get_available_types()
+            
+            assert "subprocess" in available
+            assert available["subprocess"] is True
+            assert "docker" in available
+            assert available["docker"] is True
+            assert "e2b" in available
+            assert available["e2b"] is False
+
+    def test_check_availability_subprocess(self):
+        """Test checking subprocess availability."""
+        manager = SandboxManager()
+        
+        # Subprocess should always be available
+        available = manager._check_availability("subprocess")
+        assert available is True
+
+    def test_check_availability_docker(self):
+        """Test checking Docker availability."""
+        manager = SandboxManager()
+        
+        with patch('importlib.import_module') as mock_import:
+            with patch.object(mock_import.return_value, 'DockerSandbox') as mock_class:
+                mock_instance = Mock()
+                mock_instance.is_available = True
+                mock_class.return_value = mock_instance
+                
+                available = manager._check_availability("docker")
+                assert available is True
+
+    def test_check_availability_not_available(self):
+        """Test checking availability when not available."""
+        manager = SandboxManager()
+        
+        with patch('importlib.import_module', side_effect=ImportError()):
+            available = manager._check_availability("docker")
+            assert available is False
+
+    def test_check_availability_unknown(self):
+        """Test checking availability for unknown type."""
+        manager = SandboxManager()
+        
+        available = manager._check_availability("unknown")
+        assert available is False

src/praisonai-agents/tests/unit/sandbox/test_sandbox_mixin.py

@@ -0,0 +1,177 @@
+"""
+Unit tests for SandboxMixin functionality.
+"""
+
+import pytest
+from unittest.mock import Mock, AsyncMock, patch
+from praisonaiagents.agent.sandbox_mixin import SandboxMixin
+from praisonaiagents.sandbox import SandboxConfig, SandboxResult, SandboxStatus
+
+
+class MockAgent(SandboxMixin):
+    """Mock agent class for testing SandboxMixin."""
+    
+    def __init__(self, sandbox=None, verbose=False):
+        super().__init__(sandbox=sandbox)
+        self.verbose = verbose
+
+
+class TestSandboxMixin:
+    """Test SandboxMixin functionality."""
+
+    def test_init_no_sandbox(self):
+        """Test initialization without sandbox."""
+        agent = MockAgent()
+        assert not agent.has_sandbox
+        assert agent.sandbox_config is None
+
+    def test_init_with_boolean_true(self):
+        """Test initialization with sandbox=True."""
+        agent = MockAgent(sandbox=True)
+        assert agent.has_sandbox
+        assert agent.sandbox_config is not None
+        assert agent.sandbox_config.sandbox_type == "subprocess"
+
+    def test_init_with_config(self):
+        """Test initialization with explicit config."""
+        config = SandboxConfig.docker("python:3.11")
+        agent = MockAgent(sandbox=config)
+        assert agent.has_sandbox
+        assert agent.sandbox_config == config
+
+    def test_get_sandbox_manager_no_config(self):
+        """Test get_sandbox_manager without config."""
+        agent = MockAgent()
+        manager = agent.get_sandbox_manager()
+        assert manager is None
+
+    @patch('praisonaiagents.agent.sandbox_mixin.SandboxManager')
+    def test_get_sandbox_manager_with_config(self, mock_manager_class):
+        """Test get_sandbox_manager with config."""
+        config = SandboxConfig.subprocess()
+        agent = MockAgent(sandbox=config)
+        
+        manager = agent.get_sandbox_manager()
+        
+        mock_manager_class.assert_called_once_with(config)
+        assert agent._sandbox_manager is not None
+
+    def test_execute_code_without_sandbox_raises(self):
+        """Test that execute_code raises without sandbox."""
+        agent = MockAgent()
+        
+        with pytest.raises(RuntimeError, match="No sandbox configured"):
+            agent.execute_code_sync("print('hello')")
+
+    @patch('praisonaiagents.agent.sandbox_mixin.check_code_safety')
+    @patch('praisonaiagents.agent.sandbox_mixin.SandboxManager')
+    async def test_execute_code_with_warnings(self, mock_manager_class, mock_check_safety):
+        """Test execute_code with security warnings."""
+        # Setup
+        config = SandboxConfig.subprocess()
+        agent = MockAgent(sandbox=config, verbose=True)
+        
+        mock_warnings = ["Potential security issue"]
+        mock_check_safety.return_value = mock_warnings
+        
+        mock_manager = AsyncMock()
+        mock_result = SandboxResult(status=SandboxStatus.COMPLETED, stdout="output")
+        mock_manager.run_code.return_value = mock_result
+        mock_manager_class.return_value = mock_manager
+        
+        # Execute
+        result = await agent.execute_code("print('test')", check_security=True)
+        
+        # Verify
+        mock_check_safety.assert_called_once_with("print('test')", "python")
+        mock_manager.run_code.assert_called_once()
+        assert result == mock_result
+
+    @patch('praisonaiagents.agent.sandbox_mixin.SandboxManager')
+    async def test_run_shell_command(self, mock_manager_class):
+        """Test run_shell_command functionality."""
+        config = SandboxConfig.subprocess()
+        agent = MockAgent(sandbox=config)
+        
+        # Mock manager as async context manager
+        mock_manager = AsyncMock()
+        mock_result = SandboxResult(status=SandboxStatus.COMPLETED, stdout="result")
+        mock_manager.__aenter__.return_value = mock_manager
+        mock_manager.__aexit__.return_value = None
+        mock_manager.run_command.return_value = mock_result
+        
+        mock_manager_class.return_value = mock_manager
+        
+        result = await agent.run_shell_command("echo hello")
+        
+        mock_manager.run_command.assert_called_once_with("echo hello")
+        assert result == mock_result
+
+    def test_get_sandbox_status_no_config(self):
+        """Test get_sandbox_status without config."""
+        agent = MockAgent()
+        status = agent.get_sandbox_status()
+        assert status == {"configured": False}
+
+    @patch('praisonaiagents.agent.sandbox_mixin.SandboxManager')
+    def test_get_sandbox_status_with_config(self, mock_manager_class):
+        """Test get_sandbox_status with config."""
+        config = SandboxConfig.subprocess()
+        agent = MockAgent(sandbox=config)
+        
+        mock_manager = Mock()
+        mock_manager.get_available_types.return_value = {"subprocess": True}
+        mock_manager_class.return_value = mock_manager
+        
+        status = agent.get_sandbox_status()
+        
+        assert status["configured"] is True
+        assert status["config"] is not None
+        assert status["available_types"] == {"subprocess": True}
+        assert status["current_type"] == "subprocess"
+
+    def test_get_code_execution_tools_no_sandbox(self):
+        """Test _get_code_execution_tools without sandbox."""
+        agent = MockAgent()
+        tools = agent._get_code_execution_tools()
+        assert tools == []
+
+    @patch('praisonaiagents.agent.sandbox_mixin.SandboxManager')
+    def test_get_code_execution_tools_with_sandbox(self, mock_manager_class):
+        """Test _get_code_execution_tools with sandbox."""
+        config = SandboxConfig.subprocess()
+        agent = MockAgent(sandbox=config)
+        
+        tools = agent._get_code_execution_tools()
+        
+        assert len(tools) == 2
+        # Check that tools are functions
+        assert callable(tools[0])
+        assert callable(tools[1])
+
+    async def test_sandbox_cleanup(self):
+        """Test sandbox cleanup functionality."""
+        agent = MockAgent()
+        
+        # Mock manager
+        mock_manager = AsyncMock()
+        agent._sandbox_manager = mock_manager
+        
+        await agent.sandbox_cleanup()
+        
+        mock_manager.__aexit__.assert_called_once_with(None, None, None)
+        assert agent._sandbox_manager is None
+
+    async def test_sandbox_cleanup_with_error(self):
+        """Test sandbox cleanup with error handling."""
+        agent = MockAgent()
+        
+        # Mock manager that raises exception
+        mock_manager = AsyncMock()
+        mock_manager.__aexit__.side_effect = Exception("Cleanup failed")
+        agent._sandbox_manager = mock_manager
+        
+        # Should not raise, just log warning
+        await agent.sandbox_cleanup()
+        
+        assert agent._sandbox_manager is None

src/praisonai-agents/tests/unit/sandbox/test_security.py

@@ -0,0 +1,203 @@
+"""
+Unit tests for sandbox security module.
+"""
+
+import pytest
+from praisonaiagents.sandbox.security import (
+    check_code_safety, 
+    format_warnings,
+    SecurityWarning,
+    SecurityLevel,
+    SecurityAnalyzer
+)
+
+
+class TestCodeSafety:
+    """Test code safety checking functionality."""
+
+    def test_empty_code(self):
+        """Test empty code returns no warnings."""
+        warnings = check_code_safety("", "python")
+        assert warnings == []
+
+    def test_safe_code(self):
+        """Test safe code returns no warnings."""
+        safe_code = """
+x = 1 + 2
+print(f"Result: {x}")
+"""
+        warnings = check_code_safety(safe_code, "python")
+        assert warnings == []
+
+    def test_dangerous_imports(self):
+        """Test detection of dangerous imports."""
+        dangerous_code = """
+import os
+import subprocess
+os.system('rm -rf /')
+"""
+        warnings = check_code_safety(dangerous_code, "python")
+        assert len(warnings) > 0
+        assert any("os.system" in str(w) for w in warnings)
+
+    def test_eval_exec_usage(self):
+        """Test detection of eval/exec usage."""
+        dangerous_code = """
+user_input = input("Enter code: ")
+eval(user_input)
+exec(user_input)
+"""
+        warnings = check_code_safety(dangerous_code, "python")
+        assert len(warnings) > 0
+        assert any("eval" in str(w) or "exec" in str(w) for w in warnings)
+
+    def test_network_operations(self):
+        """Test detection of network operations."""
+        network_code = """
+import urllib.request
+import socket
+urllib.request.urlopen('http://example.com')
+"""
+        warnings = check_code_safety(network_code, "python")
+        assert len(warnings) > 0
+
+    def test_file_operations(self):
+        """Test detection of file operations."""
+        file_code = """
+with open('/etc/passwd', 'r') as f:
+    data = f.read()
+"""
+        warnings = check_code_safety(file_code, "python")
+        assert len(warnings) > 0
+
+    def test_bash_commands(self):
+        """Test bash command safety checking."""
+        dangerous_bash = "rm -rf / --no-preserve-root"
+        warnings = check_code_safety(dangerous_bash, "bash")
+        assert len(warnings) > 0
+
+    def test_sql_injection_patterns(self):
+        """Test SQL injection pattern detection."""
+        sql_code = """
+query = f"SELECT * FROM users WHERE id = {user_id}"
+cursor.execute(query)
+"""
+        warnings = check_code_safety(sql_code, "python")
+        assert len(warnings) > 0
+
+    def test_format_warnings(self):
+        """Test warning formatting."""
+        warnings = [
+            SecurityWarning("Test warning", SecurityLevel.HIGH, "eval", 1),
+            SecurityWarning("Another warning", SecurityLevel.MEDIUM, "import os", 2)
+        ]
+        
+        formatted = format_warnings(warnings)
+        assert "Test warning" in formatted
+        assert "Another warning" in formatted
+        assert "HIGH" in formatted
+        assert "MEDIUM" in formatted
+
+    def test_format_warnings_empty(self):
+        """Test formatting empty warnings list."""
+        formatted = format_warnings([])
+        assert formatted == ""
+
+
+class TestSecurityAnalyzer:
+    """Test SecurityAnalyzer class."""
+
+    def test_analyzer_initialization(self):
+        """Test SecurityAnalyzer initialization."""
+        analyzer = SecurityAnalyzer()
+        assert analyzer is not None
+
+    def test_analyze_safe_code(self):
+        """Test analyzer with safe code."""
+        analyzer = SecurityAnalyzer()
+        code = "print('Hello, World!')"
+        
+        warnings = analyzer.analyze(code, "python")
+        assert warnings == []
+
+    def test_analyze_unsafe_code(self):
+        """Test analyzer with unsafe code."""
+        analyzer = SecurityAnalyzer()
+        code = "import os; os.system('rm -rf /')"
+        
+        warnings = analyzer.analyze(code, "python")
+        assert len(warnings) > 0
+        assert all(isinstance(w, SecurityWarning) for w in warnings)
+
+    def test_different_languages(self):
+        """Test analyzer with different languages."""
+        analyzer = SecurityAnalyzer()
+        
+        # Python
+        python_warnings = analyzer.analyze("import os", "python")
+        assert len(python_warnings) > 0
+        
+        # Bash
+        bash_warnings = analyzer.analyze("rm -rf /", "bash")
+        assert len(bash_warnings) > 0
+        
+        # JavaScript (should fall back to basic checks)
+        js_warnings = analyzer.analyze("eval(userInput)", "javascript")
+        assert len(js_warnings) > 0
+
+    def test_security_levels(self):
+        """Test different security levels are detected."""
+        analyzer = SecurityAnalyzer()
+        
+        # High risk
+        high_risk_code = "import os; os.system('rm -rf /')"
+        warnings = analyzer.analyze(high_risk_code, "python")
+        high_warnings = [w for w in warnings if w.level == SecurityLevel.HIGH]
+        assert len(high_warnings) > 0
+        
+        # Medium risk
+        medium_risk_code = "import subprocess"
+        warnings = analyzer.analyze(medium_risk_code, "python")
+        medium_warnings = [w for w in warnings if w.level == SecurityLevel.MEDIUM]
+        # May or may not have medium warnings depending on implementation
+
+    def test_line_numbers(self):
+        """Test that line numbers are correctly reported."""
+        analyzer = SecurityAnalyzer()
+        code = """
+print("Safe line")
+import os
+os.system('dangerous')
+"""
+        
+        warnings = analyzer.analyze(code, "python")
+        assert len(warnings) > 0
+        
+        # Check that line numbers are reasonable
+        line_numbers = [w.line_number for w in warnings if w.line_number is not None]
+        assert all(ln > 0 for ln in line_numbers)
+        assert all(ln <= 4 for ln in line_numbers)  # Within the code block
+
+    def test_pattern_matching(self):
+        """Test that specific patterns are matched correctly."""
+        analyzer = SecurityAnalyzer()
+        
+        test_cases = [
+            ("eval(x)", ["eval"]),
+            ("exec('code')", ["exec"]),
+            ("__import__('os')", ["__import__"]),
+            ("open('/etc/passwd')", ["file access"]),
+            ("subprocess.call(['rm', 'file'])", ["subprocess"]),
+        ]
+        
+        for code, expected_patterns in test_cases:
+            warnings = analyzer.analyze(code, "python")
+            assert len(warnings) > 0, f"No warnings for: {code}"
+            
+            # Check that at least one expected pattern is found
+            warning_texts = [str(w) for w in warnings]
+            found = any(
+                any(pattern.lower() in text.lower() for text in warning_texts)
+                for pattern in expected_patterns
+            )
+            assert found, f"Expected patterns {expected_patterns} not found in warnings for: {code}"

src/praisonai-agents/tests/unit/test_allowed_tools_filter.py

@@ -0,0 +1,300 @@
+"""Unit tests for the ALLOWED_TOOLS filter module."""
+import os
+import logging
+from unittest.mock import patch, MagicMock
+import pytest
+
+from praisonaiagents.allowed_tools_filter import AllowedToolsFilter, filter_tools_with_allowed_tools
+
+
+class TestAllowedToolsFilter:
+    """Test cases for AllowedToolsFilter class."""
+    
+    def test_filter_initialization_unset(self):
+        """Test filter initialization with no environment variable set."""
+        with patch.dict(os.environ, {}, clear=True):
+            filter_instance = AllowedToolsFilter()
+            assert filter_instance.env_value is None
+            assert not filter_instance.is_enabled()
+            assert filter_instance.get_whitelist() is None
+    
+    def test_filter_initialization_primary_var(self):
+        """Test filter initialization with ALLOWED_TOOLS set."""
+        with patch.dict(os.environ, {"ALLOWED_TOOLS": "search,send_message"}):
+            filter_instance = AllowedToolsFilter()
+            assert filter_instance.env_value == "search,send_message"
+            assert filter_instance.env_var_name == "ALLOWED_TOOLS"
+            assert filter_instance.is_enabled()
+            assert filter_instance.get_whitelist() == {"search", "send_message"}
+    
+    def test_filter_initialization_legacy_var(self):
+        """Test filter initialization with HERMES_ONLY_TOOLS set (backward compatibility)."""
+        with patch.dict(os.environ, {"HERMES_ONLY_TOOLS": "search,send_message"}):
+            filter_instance = AllowedToolsFilter()
+            assert filter_instance.env_value == "search,send_message"
+            assert filter_instance.env_var_name == "HERMES_ONLY_TOOLS"
+            assert filter_instance.is_enabled()
+            assert filter_instance.get_whitelist() == {"search", "send_message"}
+    
+    def test_filter_initialization_primary_precedence(self):
+        """Test that ALLOWED_TOOLS takes precedence over HERMES_ONLY_TOOLS."""
+        with patch.dict(os.environ, {
+            "ALLOWED_TOOLS": "search,send_message", 
+            "HERMES_ONLY_TOOLS": "old_tool"
+        }):
+            filter_instance = AllowedToolsFilter()
+            assert filter_instance.env_value == "search,send_message"
+            assert filter_instance.env_var_name == "ALLOWED_TOOLS"
+            assert filter_instance.get_whitelist() == {"search", "send_message"}
+    
+    def test_filter_initialization_empty_string_error(self):
+        """Test that empty string raises ValueError."""
+        with patch.dict(os.environ, {"ALLOWED_TOOLS": ""}):
+            with pytest.raises(ValueError, match="ALLOWED_TOOLS cannot be empty"):
+                AllowedToolsFilter()
+    
+    def test_filter_initialization_whitespace_only_error(self):
+        """Test that whitespace-only string raises ValueError."""
+        with patch.dict(os.environ, {"ALLOWED_TOOLS": "   "}):
+            with pytest.raises(ValueError, match="ALLOWED_TOOLS cannot be empty"):
+                AllowedToolsFilter()
+    
+    def test_filter_initialization_empty_after_split_error(self):
+        """Test that comma-only string raises ValueError."""
+        with patch.dict(os.environ, {"ALLOWED_TOOLS": ",,,"}):
+            with pytest.raises(ValueError, match="contains no valid tool names"):
+                AllowedToolsFilter()
+    
+    def test_parse_whitelist_with_spaces(self):
+        """Test parsing comma-separated values with spaces."""
+        with patch.dict(os.environ, {"ALLOWED_TOOLS": " search , send_message , extract_pdf "}):
+            filter_instance = AllowedToolsFilter()
+            assert filter_instance.get_whitelist() == {"search", "send_message", "extract_pdf"}
+    
+    @patch.dict(os.environ, {"ALLOWED_TOOLS": "search,unknown_tool,send_message", "CI": "false"})
+    def test_filter_tools_with_unknown_tools_dev_mode(self):
+        """Test filtering tools with unknown tools in development mode."""
+        filter_instance = AllowedToolsFilter()
+        available_tools = {"search", "send_message", "extract_pdf"}
+        
+        with patch.object(logging.getLogger(), 'warning') as mock_warn:
+            filtered = filter_instance.filter_tools(available_tools)
+            
+        # Should return intersection of available and whitelisted tools
+        assert filtered == {"search", "send_message"}
+        
+        # Should warn about unknown tools
+        mock_warn.assert_called()
+        call_args = mock_warn.call_args[0]
+        assert "Unknown tools" in call_args[0]
+        assert ["unknown_tool"] == call_args[2]
+    
+    @patch.dict(os.environ, {"ALLOWED_TOOLS": "search,unknown_tool,send_message", "CI": "true"})
+    def test_filter_tools_with_unknown_tools_ci_mode(self):
+        """Test filtering tools with unknown tools in CI mode."""
+        filter_instance = AllowedToolsFilter()
+        available_tools = {"search", "send_message", "extract_pdf"}
+        
+        with pytest.raises(ValueError, match="Unknown tools in ALLOWED_TOOLS"):
+            filter_instance.filter_tools(available_tools)
+    
+    @patch.dict(os.environ, {"ALLOWED_TOOLS": "search,send_message"})
+    def test_filter_tools_all_available(self):
+        """Test filtering when all whitelisted tools are available."""
+        filter_instance = AllowedToolsFilter()
+        available_tools = {"search", "send_message", "extract_pdf"}
+        
+        filtered = filter_instance.filter_tools(available_tools)
+        assert filtered == {"search", "send_message"}
+    
+    @patch.dict(os.environ, {"ALLOWED_TOOLS": "search,send_message"})
+    def test_filter_tools_partial_match(self):
+        """Test filtering when only some whitelisted tools are available."""
+        filter_instance = AllowedToolsFilter()
+        available_tools = {"search", "extract_pdf"}
+        
+        filtered = filter_instance.filter_tools(available_tools)
+        assert filtered == {"search"}
+    
+    @patch.dict(os.environ, {"ALLOWED_TOOLS": "search,send_message"})
+    def test_filter_tools_dict_input(self):
+        """Test filtering with dictionary input."""
+        filter_instance = AllowedToolsFilter()
+        available_tools = {"search": lambda: None, "send_message": lambda: None, "extract_pdf": lambda: None}
+        
+        filtered = filter_instance.filter_tools(available_tools)
+        assert filtered == {"search", "send_message"}
+    
+    @patch.dict(os.environ, {"ALLOWED_TOOLS": "search,send_message"})
+    def test_filter_tools_list_input(self):
+        """Test filtering with list input."""
+        filter_instance = AllowedToolsFilter()
+        available_tools = ["search", "send_message", "extract_pdf"]
+        
+        filtered = filter_instance.filter_tools(available_tools)
+        assert filtered == {"search", "send_message"}
+    
+    @patch.dict(os.environ, {"ALLOWED_TOOLS": "search,send_message,unknown_in_whitelist"})
+    def test_diagnostics_data(self):
+        """Test diagnostics data collection."""
+        filter_instance = AllowedToolsFilter()
+        available_tools = {"search", "send_message", "extract_pdf"}
+        
+        filtered = filter_instance.filter_tools(available_tools)
+        diagnostics = filter_instance.get_diagnostics()
+        
+        assert diagnostics["env_var_name"] == "ALLOWED_TOOLS"
+        assert diagnostics["env_value"] == "search,send_message,unknown_in_whitelist"
+        assert diagnostics["whitelist"] == ["search", "send_message", "unknown_in_whitelist"]
+        assert set(diagnostics["registered_before_filter"]) == {"search", "send_message", "extract_pdf"}
+        assert set(diagnostics["registered_after_filter"]) == {"search", "send_message"}
+        assert set(diagnostics["dropped_tools"]) == {"extract_pdf"}
+        assert set(diagnostics["unknown_tools"]) == {"unknown_in_whitelist"}
+    
+    @patch.dict(os.environ, {"ALLOWED_TOOLS": "search,send_message"})
+    def test_log_diagnostics(self, caplog):
+        """Test diagnostics logging."""
+        filter_instance = AllowedToolsFilter()
+        available_tools = {"search", "send_message", "extract_pdf"}
+        
+        with caplog.at_level(logging.INFO):
+            filter_instance.filter_tools(available_tools)
+            filter_instance.log_diagnostics()
+        
+        log_text = caplog.text
+        assert "ALLOWED_TOOLS FILTER DIAGNOSTICS" in log_text
+        assert "ALLOWED_TOOLS=search,send_message" in log_text
+        assert "RegisteredBeforeFilter" in log_text
+        assert "RegisteredAfterFilter" in log_text
+        assert "DroppedTools" in log_text
+        assert "UnknownTools" in log_text
+    
+    def test_log_diagnostics_before_filtering(self, caplog):
+        """Test logging diagnostics before calling filter_tools."""
+        filter_instance = AllowedToolsFilter()
+        
+        with caplog.at_level(logging.WARNING):
+            filter_instance.log_diagnostics()
+        
+        assert "No diagnostics available" in caplog.text
+    
+    def test_filter_tools_unset_env_var(self, caplog):
+        """Test filtering with unset environment variable."""
+        with patch.dict(os.environ, {}, clear=True):
+            filter_instance = AllowedToolsFilter()
+            available_tools = {"search", "send_message", "extract_pdf"}
+            
+            with caplog.at_level(logging.WARNING):
+                filtered = filter_instance.filter_tools(available_tools)
+            
+            # Should return all tools
+            assert filtered == available_tools
+            assert "Tool whitelist is unset" in caplog.text
+    
+    @patch.dict(os.environ, {"ALLOWED_TOOLS": "tool1", "CI": "false"})
+    def test_no_matching_tools(self):
+        """Test filtering when no whitelisted tools match available tools."""
+        filter_instance = AllowedToolsFilter()
+        available_tools = {"search", "send_message", "extract_pdf"}
+        
+        filtered = filter_instance.filter_tools(available_tools)
+        assert filtered == set()  # Empty set when no matches
+    
+    def test_custom_env_var_name(self):
+        """Test using custom environment variable name."""
+        with patch.dict(os.environ, {"CUSTOM_TOOLS": "search,send_message"}):
+            filter_instance = AllowedToolsFilter("CUSTOM_TOOLS")
+            assert filter_instance.env_var_name == "CUSTOM_TOOLS"
+            assert filter_instance.get_whitelist() == {"search", "send_message"}
+
+
+class TestConvenienceFunction:
+    """Test cases for the convenience function."""
+    
+    @patch.dict(os.environ, {"ALLOWED_TOOLS": "search,send_message"})
+    def test_filter_tools_with_allowed_tools(self):
+        """Test the convenience function."""
+        available_tools = {"search", "send_message", "extract_pdf"}
+        
+        filtered = filter_tools_with_allowed_tools(available_tools, log_diagnostics=False)
+        assert filtered == {"search", "send_message"}
+    
+    @patch.dict(os.environ, {"CUSTOM_TOOLS": "search"})
+    def test_filter_tools_with_custom_env_var(self):
+        """Test the convenience function with custom env var."""
+        available_tools = {"search", "send_message"}
+        
+        filtered = filter_tools_with_allowed_tools(available_tools, env_var_name="CUSTOM_TOOLS", log_diagnostics=False)
+        assert filtered == {"search"}
+    
+    @patch.dict(os.environ, {"ALLOWED_TOOLS": "search"})
+    def test_filter_tools_with_diagnostics(self, caplog):
+        """Test the convenience function with diagnostics enabled."""
+        available_tools = {"search", "send_message"}
+        
+        with caplog.at_level(logging.INFO):
+            filtered = filter_tools_with_allowed_tools(available_tools, log_diagnostics=True)
+        
+        assert filtered == {"search"}
+        assert "ALLOWED_TOOLS FILTER DIAGNOSTICS" in caplog.text
+
+
+class TestBackwardCompatibility:
+    """Test backward compatibility aliases."""
+    
+    @patch.dict(os.environ, {"HERMES_ONLY_TOOLS": "search,send_message"})
+    def test_legacy_class_alias(self):
+        """Test HermesToolFilter alias works."""
+        from praisonaiagents.allowed_tools_filter import HermesToolFilter
+        
+        filter_instance = HermesToolFilter()
+        assert isinstance(filter_instance, AllowedToolsFilter)
+        assert filter_instance.env_var_name == "HERMES_ONLY_TOOLS"
+        assert filter_instance.get_whitelist() == {"search", "send_message"}
+    
+    @patch.dict(os.environ, {"HERMES_ONLY_TOOLS": "search,send_message"})
+    def test_legacy_function_aliases(self):
+        """Test legacy function aliases work."""
+        from praisonaiagents.allowed_tools_filter import (
+            filter_tools_with_hermes, hermes_filter, apply_hermes_filter
+        )
+        
+        available_tools = {"search", "send_message", "extract_pdf"}
+        
+        # All aliases should work and return the same result
+        result1 = filter_tools_with_hermes(available_tools, log_diagnostics=False)
+        result2 = hermes_filter(available_tools, log_diagnostics=False)
+        result3 = apply_hermes_filter(available_tools, log_diagnostics=False)
+        
+        expected = {"search", "send_message"}
+        assert result1 == expected
+        assert result2 == expected
+        assert result3 == expected
+    
+    def test_empty_tools_edge_case(self):
+        """Test filtering with empty available tools."""
+        with patch.dict(os.environ, {"ALLOWED_TOOLS": "search,send_message"}):
+            filter_instance = AllowedToolsFilter()
+            
+            filtered = filter_instance.filter_tools(set())
+            assert filtered == set()
+            
+            # Also test with empty list and dict
+            assert filter_instance.filter_tools([]) == set()
+            assert filter_instance.filter_tools({}) == set()
+    
+    def test_whitelist_immutability(self):
+        """Test that get_whitelist returns a copy, not the original."""
+        with patch.dict(os.environ, {"ALLOWED_TOOLS": "search,send_message"}):
+            filter_instance = AllowedToolsFilter()
+            
+            whitelist1 = filter_instance.get_whitelist()
+            whitelist2 = filter_instance.get_whitelist()
+            
+            # Should be equal but not the same object
+            assert whitelist1 == whitelist2
+            assert whitelist1 is not whitelist2
+            
+            # Modifying one shouldn't affect the other
+            whitelist1.add("new_tool")
+            assert whitelist1 != whitelist2

src/praisonai-agents/tests/unit/test_botos_integration.py

@@ -21,6 +21,16 @@ import pytest
 from unittest.mock import MagicMock
 
 
+def _unregister_bot_platform(name: str) -> None:
+    """Remove a runtime-registered bot platform (tests only)."""
+    from praisonai.bots._registry import get_default_bot_registry
+
+    try:
+        get_default_bot_registry().unregister(name.lower())
+    except (KeyError, ValueError):
+        pass
+
+
 # ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
 # Helper: Mock adapter that simulates a platform bot
 # ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
@@ -59,9 +69,10 @@ class TestBotWithMockAdapter:
     """Integration tests for Bot class with a mock platform adapter."""
 
     def setup_method(self):
-        from praisonai.bots._registry import register_platform, _custom_platforms
+        from praisonai.bots._registry import register_platform
+
         register_platform("mockbot", MockPlatformAdapter)
-        self._cleanup = lambda: _custom_platforms.pop("mockbot", None)
+        self._cleanup = lambda: _unregister_bot_platform("mockbot")
 
     def teardown_method(self):
         self._cleanup()
@@ -148,10 +159,14 @@ class TestBotOSLifecycle:
     """Integration tests for BotOS multi-bot orchestration."""
 
     def setup_method(self):
-        from praisonai.bots._registry import register_platform, _custom_platforms
+        from praisonai.bots._registry import register_platform
+
         register_platform("mock1", MockPlatformAdapter)
         register_platform("mock2", MockPlatformAdapter)
-        self._cleanup = lambda: (_custom_platforms.pop("mock1", None), _custom_platforms.pop("mock2", None))
+        self._cleanup = lambda: (
+            _unregister_bot_platform("mock1"),
+            _unregister_bot_platform("mock2"),
+        )
 
     def teardown_method(self):
         self._cleanup()
@@ -212,9 +227,10 @@ class TestRealAgentInBot:
     """Tests with a real praisonaiagents.Agent inside Bot."""
 
     def setup_method(self):
-        from praisonai.bots._registry import register_platform, _custom_platforms
+        from praisonai.bots._registry import register_platform
+
         register_platform("mockbot", MockPlatformAdapter)
-        self._cleanup = lambda: _custom_platforms.pop("mockbot", None)
+        self._cleanup = lambda: _unregister_bot_platform("mockbot")
 
     def teardown_method(self):
         self._cleanup()
@@ -246,9 +262,10 @@ class TestAgentTeamInBotOS:
     """AgentTeam (multi-agent) inside Bot and BotOS."""
 
     def setup_method(self):
-        from praisonai.bots._registry import register_platform, _custom_platforms
+        from praisonai.bots._registry import register_platform
+
         register_platform("mockbot", MockPlatformAdapter)
-        self._cleanup = lambda: _custom_platforms.pop("mockbot", None)
+        self._cleanup = lambda: _unregister_bot_platform("mockbot")
 
     def teardown_method(self):
         self._cleanup()
@@ -300,9 +317,10 @@ class TestAgentFlowInBotOS:
     """AgentFlow (deterministic workflow) inside Bot and BotOS."""
 
     def setup_method(self):
-        from praisonai.bots._registry import register_platform, _custom_platforms
+        from praisonai.bots._registry import register_platform
+
         register_platform("mockbot", MockPlatformAdapter)
-        self._cleanup = lambda: _custom_platforms.pop("mockbot", None)
+        self._cleanup = lambda: _unregister_bot_platform("mockbot")
 
     def teardown_method(self):
         self._cleanup()
@@ -355,9 +373,8 @@ class TestPlatformExtensibility:
     """Third-party platform registration and usage."""
 
     def teardown_method(self):
-        from praisonai.bots._registry import _custom_platforms
-        _custom_platforms.pop("custom_chat", None)
-        _custom_platforms.pop("my_platform", None)
+        _unregister_bot_platform("custom_chat")
+        _unregister_bot_platform("my_platform")
 
     def test_register_and_use_custom_platform(self):
         from praisonai.bots._registry import register_platform, resolve_adapter
@@ -389,7 +406,7 @@ class TestPlatformExtensibility:
 
     def test_resolve_unknown_platform_raises(self):
         from praisonai.bots._registry import resolve_adapter
-        with pytest.raises(ValueError, match="Unknown platform"):
+        with pytest.raises(ValueError, match=r"Unknown .*plugin"):
             resolve_adapter("nonexistent_xyz")
 
 
@@ -401,9 +418,10 @@ class TestBotOSFromConfig:
     """BotOS.from_config() YAML config loading."""
 
     def setup_method(self):
-        from praisonai.bots._registry import register_platform, _custom_platforms
+        from praisonai.bots._registry import register_platform
+
         register_platform("mockbot", MockPlatformAdapter)
-        self._cleanup = lambda: _custom_platforms.pop("mockbot", None)
+        self._cleanup = lambda: _unregister_bot_platform("mockbot")
 
     def teardown_method(self):
         self._cleanup()
@@ -459,7 +477,8 @@ platforms:
 
     def test_from_config_multi_platform(self):
         from praisonai.bots import BotOS
-        from praisonai.bots._registry import register_platform, _custom_platforms
+        from praisonai.bots._registry import register_platform
+
         register_platform("mock2", MockPlatformAdapter)
 
         yaml_content = """\
@@ -484,7 +503,7 @@ platforms:
             assert "mock2" in botos.list_bots()
         finally:
             os.unlink(path)
-            _custom_platforms.pop("mock2", None)
+            _unregister_bot_platform("mock2")
 
     def test_from_config_invalid_raises(self):
         from praisonai.bots import BotOS

src/praisonai-agents/tests/unit/tools/test_mentions_url_ssrf.py

@@ -0,0 +1,10 @@
+"""@url mentions must not fetch loopback targets."""
+
+from praisonaiagents.tools.mentions import MentionsParser
+
+
+def test_url_mention_blocks_loopback():
+    parser = MentionsParser()
+    result = parser._process_url_mention("http://127.0.0.1:8765/")
+    assert result is not None
+    assert "Blocked" in result

src/praisonai-agents/tests/unit/tools/test_profiles.py

@@ -361,3 +361,11 @@ class TestProfileExports:
         from praisonaiagents.tools import AUTONOMY_PROFILE
         
         assert AUTONOMY_PROFILE is not None
+
+    def test_import_execute_code_uses_module_function(self):
+        """execute_code should resolve to python_tools.execute_code function."""
+        from praisonaiagents.tools import execute_code
+        from praisonaiagents.tools.python_tools import execute_code as execute_code_impl
+
+        assert callable(execute_code)
+        assert execute_code is execute_code_impl

src/praisonai-agents/tests/unit/tools/test_python_tools_sandbox.py

@@ -112,6 +112,22 @@ class TestSandboxEscapePrevention:
         result = sandbox.run("__import__('os')")
         assert result["success"] is False
 
+    def test_print_self_blocked(self, sandbox):
+        """print.__self__ leaks real builtins module (GHSA-4mr5-g6f9-cfrh)."""
+        result = sandbox.run("b = print.__self__")
+        assert result["success"] is False
+        assert "restricted" in result["stderr"].lower()
+
+    def test_vars_call_blocked(self, sandbox):
+        """vars() can expose builtins.__dict__ after __self__ leak."""
+        result = sandbox.run("vars({})")
+        assert result["success"] is False
+
+    def test_attribute_dunder_call_blocked(self, sandbox):
+        """Attribute calls bypass bare-name Call checks."""
+        result = sandbox.run("(1).__class__.__mro__")
+        assert result["success"] is False
+
 
 # ── Legitimate Code (all must PASS) ─────────────────────────────────────────
 

src/praisonai-agents/tests/unit/tools/test_spider_url_validation.py

@@ -38,6 +38,16 @@ def test_still_blocks_loopback():
     assert spider._validate_url("http://localhost/") is False
 
 
+def test_blocks_alternate_loopback_encodings():
+    """GHSA-5c6w-wwfq-7qqm: non-canonical loopback host forms."""
+    spider = SpiderTools()
+    assert spider._validate_url("http://localhost.:8765/") is False
+    assert spider._validate_url("http://127.1:8765/") is False
+    assert spider._validate_url("http://0177.0.0.1:8765/") is False
+    assert spider._validate_url("http://0x7f000001:8765/") is False
+    assert spider._validate_url("http://2130706433:8765/") is False
+
+
 def test_rejects_non_string_input():
     spider = SpiderTools()
     assert spider._validate_url(None) is False  # type: ignore[arg-type]

src/praisonai-platform/praisonai_platform/__main__.py

@@ -7,12 +7,18 @@ Usage::
 """
 
 import argparse
+import os
 import sys
 
 
 def main() -> None:
+    default_host = os.environ.get("PLATFORM_HOST", "127.0.0.1")
     parser = argparse.ArgumentParser(description="PraisonAI Platform Server")
-    parser.add_argument("--host", default="0.0.0.0", help="Bind host (default: 0.0.0.0)")
+    parser.add_argument(
+        "--host",
+        default=default_host,
+        help="Bind host (default: 127.0.0.1, or PLATFORM_HOST env)",
+    )
     parser.add_argument("--port", type=int, default=8000, help="Bind port (default: 8000)")
     parser.add_argument("--reload", action="store_true", help="Enable auto-reload for development")
     args = parser.parse_args()

src/praisonai-platform/praisonai_platform/api/deps.py

@@ -71,3 +71,57 @@ async def require_workspace_member(
         )
     user.workspace_id = workspace_id
     return user
+
+
+async def require_workspace_admin(
+    workspace_id: str,
+    user: AuthIdentity = Depends(get_current_user),
+    session: AsyncSession = Depends(get_db),
+) -> AuthIdentity:
+    """Require admin or owner role in the workspace."""
+    return await require_workspace_member(
+        workspace_id, user, session, min_role="admin"
+    )
+
+
+async def require_workspace_owner(
+    workspace_id: str,
+    user: AuthIdentity = Depends(get_current_user),
+    session: AsyncSession = Depends(get_db),
+) -> AuthIdentity:
+    """Require owner role in the workspace."""
+    return await require_workspace_member(
+        workspace_id, user, session, min_role="owner"
+    )
+
+
+def ensure_resource_in_workspace(
+    resource_workspace_id: str | None,
+    workspace_id: str,
+    *,
+    label: str = "Resource",
+) -> None:
+    """Reject cross-workspace access (IDOR) with a generic 404."""
+    if resource_workspace_id != workspace_id:
+        raise HTTPException(
+            status_code=status.HTTP_404_NOT_FOUND,
+            detail=f"{label} not found",
+        )
+
+
+async def require_issue_in_workspace(
+    workspace_id: str,
+    issue_id: str,
+    session: AsyncSession,
+):
+    """Load an issue and verify it belongs to the URL workspace."""
+    from ..db.models import Issue
+
+    issue = await session.get(Issue, issue_id)
+    if issue is None:
+        raise HTTPException(
+            status_code=status.HTTP_404_NOT_FOUND,
+            detail="Issue not found",
+        )
+    ensure_resource_in_workspace(issue.workspace_id, workspace_id, label="Issue")
+    return issue

src/praisonai-platform/praisonai_platform/api/routes/activity.py

@@ -9,7 +9,7 @@ from sqlalchemy.ext.asyncio import AsyncSession
 
 from praisonaiagents.auth import AuthIdentity
 
-from ..deps import get_db, require_workspace_member
+from ..deps import get_db, require_issue_in_workspace, require_workspace_member
 from ..schemas import ActivityLogResponse
 from ...services.activity_service import ActivityService
 
@@ -38,6 +38,7 @@ async def list_issue_activity(
     user: AuthIdentity = Depends(require_workspace_member),
     session: AsyncSession = Depends(get_db),
 ):
+    await require_issue_in_workspace(workspace_id, issue_id, session)
     svc = ActivityService(session)
     logs = await svc.list_for_issue(issue_id, limit=limit, offset=offset)
     return [ActivityLogResponse.model_validate(log) for log in logs]

src/praisonai-platform/praisonai_platform/api/routes/agents.py

@@ -58,7 +58,7 @@ async def get_agent(
     session: AsyncSession = Depends(get_db),
 ):
     svc = AgentService(session)
-    agent = await svc.get(agent_id)
+    agent = await svc.get(agent_id, workspace_id=workspace_id)
     if agent is None:
         raise HTTPException(status_code=404, detail="Agent not found")
     return AgentResponse.model_validate(agent)
@@ -75,6 +75,7 @@ async def update_agent(
     svc = AgentService(session)
     agent = await svc.update(
         agent_id,
+        workspace_id=workspace_id,
         name=body.name,
         status=body.status,
         instructions=body.instructions,
@@ -95,6 +96,6 @@ async def delete_agent(
     session: AsyncSession = Depends(get_db),
 ):
     svc = AgentService(session)
-    deleted = await svc.delete(agent_id)
+    deleted = await svc.delete(agent_id, workspace_id=workspace_id)
     if not deleted:
         raise HTTPException(status_code=404, detail="Agent not found")

src/praisonai-platform/praisonai_platform/api/routes/dependencies.py

@@ -9,9 +9,9 @@ from sqlalchemy.ext.asyncio import AsyncSession
 
 from praisonaiagents.auth import AuthIdentity
 
-from ..deps import get_db, require_workspace_member
-from ..schemas import DependencyCreate, DependencyResponse
+from ..deps import get_db, require_issue_in_workspace, require_workspace_member
 from ...services.dependency_service import DependencyService
+from ..schemas import DependencyCreate, DependencyResponse
 
 router = APIRouter(
     prefix="/workspaces/{workspace_id}/issues/{issue_id}/dependencies",
@@ -27,6 +27,10 @@ async def create_dependency(
     user: AuthIdentity = Depends(require_workspace_member),
     session: AsyncSession = Depends(get_db),
 ):
+    await require_issue_in_workspace(workspace_id, issue_id, session)
+    await require_issue_in_workspace(
+        workspace_id, body.depends_on_issue_id, session
+    )
     svc = DependencyService(session)
     dep = await svc.create(issue_id, body.depends_on_issue_id, body.type)
     return DependencyResponse.model_validate(dep)
@@ -39,6 +43,7 @@ async def list_dependencies(
     user: AuthIdentity = Depends(require_workspace_member),
     session: AsyncSession = Depends(get_db),
 ):
+    await require_issue_in_workspace(workspace_id, issue_id, session)
     svc = DependencyService(session)
     deps = await svc.list_for_issue(issue_id)
     return [DependencyResponse.model_validate(d) for d in deps]
@@ -52,7 +57,13 @@ async def delete_dependency(
     user: AuthIdentity = Depends(require_workspace_member),
     session: AsyncSession = Depends(get_db),
 ):
+    await require_issue_in_workspace(workspace_id, issue_id, session)
     svc = DependencyService(session)
+    dep = await svc.get(dep_id)
+    if dep is None:
+        raise HTTPException(status_code=404, detail="Dependency not found")
+    if dep.issue_id != issue_id and dep.depends_on_issue_id != issue_id:
+        raise HTTPException(status_code=404, detail="Dependency not found")
     deleted = await svc.delete(dep_id)
     if not deleted:
         raise HTTPException(status_code=404, detail="Dependency not found")

src/praisonai-platform/praisonai_platform/api/routes/issues.py

@@ -9,7 +9,7 @@ from sqlalchemy.ext.asyncio import AsyncSession
 
 from praisonaiagents.auth import AuthIdentity
 
-from ..deps import get_db, require_workspace_member
+from ..deps import ensure_resource_in_workspace, get_db, require_workspace_member
 from ..schemas import (
     CommentCreate,
     CommentResponse,
@@ -87,7 +87,7 @@ async def get_issue(
     session: AsyncSession = Depends(get_db),
 ):
     svc = IssueService(session)
-    issue = await svc.get(issue_id)
+    issue = await svc.get(issue_id, workspace_id=workspace_id)
     if issue is None:
         raise HTTPException(status_code=404, detail="Issue not found")
     return IssueResponse.model_validate(issue)
@@ -104,6 +104,7 @@ async def update_issue(
     svc = IssueService(session)
     issue = await svc.update(
         issue_id,
+        workspace_id=workspace_id,
         title=body.title,
         description=body.description,
         status=body.status,
@@ -132,7 +133,7 @@ async def delete_issue(
     session: AsyncSession = Depends(get_db),
 ):
     svc = IssueService(session)
-    deleted = await svc.delete(issue_id)
+    deleted = await svc.delete(issue_id, workspace_id=workspace_id)
     if not deleted:
         raise HTTPException(status_code=404, detail="Issue not found")
 
@@ -148,6 +149,11 @@ async def add_comment(
     user: AuthIdentity = Depends(require_workspace_member),
     session: AsyncSession = Depends(get_db),
 ):
+    issue_svc = IssueService(session)
+    issue = await issue_svc.get(issue_id)
+    if issue is None:
+        raise HTTPException(status_code=404, detail="Issue not found")
+    ensure_resource_in_workspace(issue.workspace_id, workspace_id, label="Issue")
     svc = CommentService(session)
     comment = await svc.create(
         issue_id=issue_id,
@@ -166,6 +172,11 @@ async def list_comments(
     user: AuthIdentity = Depends(require_workspace_member),
     session: AsyncSession = Depends(get_db),
 ):
+    issue_svc = IssueService(session)
+    issue = await issue_svc.get(issue_id)
+    if issue is None:
+        raise HTTPException(status_code=404, detail="Issue not found")
+    ensure_resource_in_workspace(issue.workspace_id, workspace_id, label="Issue")
     svc = CommentService(session)
     comments = await svc.list_for_issue(issue_id)
     return [CommentResponse.model_validate(c) for c in comments]

src/praisonai-platform/praisonai_platform/api/routes/labels.py

@@ -9,7 +9,12 @@ from sqlalchemy.ext.asyncio import AsyncSession
 
 from praisonaiagents.auth import AuthIdentity
 
-from ..deps import get_db, require_workspace_member
+from ..deps import (
+    ensure_resource_in_workspace,
+    get_db,
+    require_issue_in_workspace,
+    require_workspace_member,
+)
 from ..schemas import LabelCreate, LabelResponse, LabelUpdate
 from ...services.label_service import LabelService
 
@@ -48,6 +53,10 @@ async def update_label(
     session: AsyncSession = Depends(get_db),
 ):
     svc = LabelService(session)
+    label = await svc.get(label_id)
+    if label is None:
+        raise HTTPException(status_code=404, detail="Label not found")
+    ensure_resource_in_workspace(label.workspace_id, workspace_id, label="Label")
     label = await svc.update(label_id, body.name, body.color)
     if label is None:
         raise HTTPException(status_code=404, detail="Label not found")
@@ -62,6 +71,10 @@ async def delete_label(
     session: AsyncSession = Depends(get_db),
 ):
     svc = LabelService(session)
+    label = await svc.get(label_id)
+    if label is None:
+        raise HTTPException(status_code=404, detail="Label not found")
+    ensure_resource_in_workspace(label.workspace_id, workspace_id, label="Label")
     deleted = await svc.delete(label_id)
     if not deleted:
         raise HTTPException(status_code=404, detail="Label not found")
@@ -78,7 +91,12 @@ async def add_label_to_issue(
     user: AuthIdentity = Depends(require_workspace_member),
     session: AsyncSession = Depends(get_db),
 ):
+    await require_issue_in_workspace(workspace_id, issue_id, session)
     svc = LabelService(session)
+    label = await svc.get(label_id)
+    if label is None:
+        raise HTTPException(status_code=404, detail="Label not found")
+    ensure_resource_in_workspace(label.workspace_id, workspace_id, label="Label")
     await svc.add_to_issue(issue_id, label_id)
 
 
@@ -90,7 +108,12 @@ async def remove_label_from_issue(
     user: AuthIdentity = Depends(require_workspace_member),
     session: AsyncSession = Depends(get_db),
 ):
+    await require_issue_in_workspace(workspace_id, issue_id, session)
     svc = LabelService(session)
+    label = await svc.get(label_id)
+    if label is None:
+        raise HTTPException(status_code=404, detail="Label not found")
+    ensure_resource_in_workspace(label.workspace_id, workspace_id, label="Label")
     await svc.remove_from_issue(issue_id, label_id)
 
 
@@ -101,6 +124,7 @@ async def list_issue_labels(
     user: AuthIdentity = Depends(require_workspace_member),
     session: AsyncSession = Depends(get_db),
 ):
+    await require_issue_in_workspace(workspace_id, issue_id, session)
     svc = LabelService(session)
     labels = await svc.list_for_issue(issue_id)
     return [LabelResponse.model_validate(l) for l in labels]

src/praisonai-platform/praisonai_platform/api/routes/projects.py

@@ -56,7 +56,7 @@ async def get_project(
     session: AsyncSession = Depends(get_db),
 ):
     svc = ProjectService(session)
-    project = await svc.get(project_id)
+    project = await svc.get(project_id, workspace_id=workspace_id)
     if project is None:
         raise HTTPException(status_code=404, detail="Project not found")
     return ProjectResponse.model_validate(project)
@@ -73,6 +73,7 @@ async def update_project(
     svc = ProjectService(session)
     project = await svc.update(
         project_id,
+        workspace_id=workspace_id,
         title=body.title,
         description=body.description,
         status=body.status,
@@ -92,7 +93,7 @@ async def delete_project(
     session: AsyncSession = Depends(get_db),
 ):
     svc = ProjectService(session)
-    deleted = await svc.delete(project_id)
+    deleted = await svc.delete(project_id, workspace_id=workspace_id)
     if not deleted:
         raise HTTPException(status_code=404, detail="Project not found")
 
@@ -105,4 +106,7 @@ async def project_stats(
     session: AsyncSession = Depends(get_db),
 ):
     svc = ProjectService(session)
+    project = await svc.get(project_id, workspace_id=workspace_id)
+    if project is None:
+        raise HTTPException(status_code=404, detail="Project not found")
     return await svc.get_stats(project_id)

src/praisonai-platform/praisonai_platform/api/routes/workspaces.py

@@ -9,7 +9,13 @@ from sqlalchemy.ext.asyncio import AsyncSession
 
 from praisonaiagents.auth import AuthIdentity
 
-from ..deps import get_current_user, get_db, require_workspace_member
+from ..deps import (
+    get_current_user,
+    get_db,
+    require_workspace_admin,
+    require_workspace_member,
+    require_workspace_owner,
+)
 from ..schemas import (
     MemberAdd,
     MemberResponse,
@@ -64,7 +70,7 @@ async def get_workspace(
 async def update_workspace(
     workspace_id: str,
     body: WorkspaceUpdate,
-    user: AuthIdentity = Depends(require_workspace_member),
+    user: AuthIdentity = Depends(require_workspace_admin),
     session: AsyncSession = Depends(get_db),
 ):
     ws_svc = WorkspaceService(session)
@@ -77,7 +83,7 @@ async def update_workspace(
 @router.delete("/{workspace_id}", status_code=status.HTTP_204_NO_CONTENT)
 async def delete_workspace(
     workspace_id: str,
-    user: AuthIdentity = Depends(require_workspace_member),
+    user: AuthIdentity = Depends(require_workspace_owner),
     session: AsyncSession = Depends(get_db),
 ):
     ws_svc = WorkspaceService(session)
@@ -93,10 +99,16 @@ async def delete_workspace(
 async def add_member(
     workspace_id: str,
     body: MemberAdd,
-    user: AuthIdentity = Depends(require_workspace_member),
+    user: AuthIdentity = Depends(require_workspace_admin),
     session: AsyncSession = Depends(get_db),
 ):
     member_svc = MemberService(session)
+    if body.role in ("owner", "admin"):
+        if not await member_svc.has_role(workspace_id, user.id, "owner"):
+            raise HTTPException(
+                status_code=status.HTTP_403_FORBIDDEN,
+                detail="Only owners can add admin or owner roles",
+            )
     member = await member_svc.add(workspace_id, body.user_id, body.role)
     return MemberResponse.model_validate(member)
 
@@ -117,10 +129,32 @@ async def update_member_role(
     workspace_id: str,
     user_id: str,
     body: MemberUpdate,
-    user: AuthIdentity = Depends(require_workspace_member),
+    user: AuthIdentity = Depends(require_workspace_admin),
     session: AsyncSession = Depends(get_db),
 ):
     member_svc = MemberService(session)
+    target = await member_svc.get(workspace_id, user_id)
+    if target is None:
+        raise HTTPException(status_code=404, detail="Member not found")
+    if user_id == user.id and body.role != target.role:
+        raise HTTPException(
+            status_code=status.HTTP_403_FORBIDDEN,
+            detail="Cannot change your own role",
+        )
+    if body.role in ("owner", "admin") and not await member_svc.has_role(
+        workspace_id, user.id, "owner"
+    ):
+        raise HTTPException(
+            status_code=status.HTTP_403_FORBIDDEN,
+            detail="Only owners can assign admin or owner roles",
+        )
+    if target.role == "owner" and not await member_svc.has_role(
+        workspace_id, user.id, "owner"
+    ):
+        raise HTTPException(
+            status_code=status.HTTP_403_FORBIDDEN,
+            detail="Only owners can change an owner's role",
+        )
     member = await member_svc.update_role(workspace_id, user_id, body.role)
     if member is None:
         raise HTTPException(status_code=404, detail="Member not found")
@@ -131,10 +165,25 @@ async def update_member_role(
 async def remove_member(
     workspace_id: str,
     user_id: str,
-    user: AuthIdentity = Depends(require_workspace_member),
+    user: AuthIdentity = Depends(require_workspace_admin),
     session: AsyncSession = Depends(get_db),
 ):
     member_svc = MemberService(session)
+    if user_id == user.id:
+        raise HTTPException(
+            status_code=status.HTTP_403_FORBIDDEN,
+            detail="Cannot remove yourself from the workspace",
+        )
+    target = await member_svc.get(workspace_id, user_id)
+    if target is None:
+        raise HTTPException(status_code=404, detail="Member not found")
+    if target.role == "owner" and not await member_svc.has_role(
+        workspace_id, user.id, "owner"
+    ):
+        raise HTTPException(
+            status_code=status.HTTP_403_FORBIDDEN,
+            detail="Only owners can remove an owner",
+        )
     removed = await member_svc.remove(workspace_id, user_id)
     if not removed:
         raise HTTPException(status_code=404, detail="Member not found")

src/praisonai-platform/praisonai_platform/services/agent_service.py

@@ -50,9 +50,16 @@ class AgentService:
         await self._session.flush()
         return agent
 
-    async def get(self, agent_id: str) -> Optional[Agent]:
-        """Get agent by ID."""
-        return await self._session.get(Agent, agent_id)
+    async def get(
+        self, agent_id: str, *, workspace_id: Optional[str] = None
+    ) -> Optional[Agent]:
+        """Get agent by ID, optionally scoped to a workspace."""
+        agent = await self._session.get(Agent, agent_id)
+        if agent is None:
+            return None
+        if workspace_id is not None and agent.workspace_id != workspace_id:
+            return None
+        return agent
 
     async def list_for_workspace(
         self,
@@ -72,6 +79,8 @@ class AgentService:
     async def update(
         self,
         agent_id: str,
+        *,
+        workspace_id: Optional[str] = None,
         name: Optional[str] = None,
         status: Optional[str] = None,
         instructions: Optional[str] = None,
@@ -80,7 +89,7 @@ class AgentService:
         max_concurrent_tasks: Optional[int] = None,
     ) -> Optional[Agent]:
         """Update agent fields."""
-        agent = await self.get(agent_id)
+        agent = await self.get(agent_id, workspace_id=workspace_id)
         if agent is None:
             return None
         if name is not None:
@@ -102,9 +111,11 @@ class AgentService:
         await self._session.flush()
         return agent
 
-    async def delete(self, agent_id: str) -> bool:
+    async def delete(
+        self, agent_id: str, *, workspace_id: Optional[str] = None
+    ) -> bool:
         """Delete an agent."""
-        agent = await self.get(agent_id)
+        agent = await self.get(agent_id, workspace_id=workspace_id)
         if agent is None:
             return False
         await self._session.delete(agent)

src/praisonai-platform/praisonai_platform/services/auth_service.py

@@ -113,6 +113,10 @@ class AuthService:
 
     def _issue_token(self, user: User) -> str:
         """Issue a JWT for a user."""
+        if JWT_SECRET == _DEFAULT_SECRET and os.environ.get("PLATFORM_ENV", "dev") != "dev":
+            raise RuntimeError(
+                "Refusing to issue JWT with default PLATFORM_JWT_SECRET outside dev"
+            )
         now = datetime.now(timezone.utc)
         payload = {
             "sub": user.id,

src/praisonai-platform/praisonai_platform/services/issue_service.py

@@ -69,9 +69,16 @@ class IssueService:
         await self._session.flush()
         return issue
 
-    async def get(self, issue_id: str) -> Optional[Issue]:
-        """Get issue by ID."""
-        return await self._session.get(Issue, issue_id)
+    async def get(
+        self, issue_id: str, *, workspace_id: Optional[str] = None
+    ) -> Optional[Issue]:
+        """Get issue by ID, optionally scoped to a workspace."""
+        issue = await self._session.get(Issue, issue_id)
+        if issue is None:
+            return None
+        if workspace_id is not None and issue.workspace_id != workspace_id:
+            return None
+        return issue
 
     async def list_for_workspace(
         self,
@@ -97,6 +104,8 @@ class IssueService:
     async def update(
         self,
         issue_id: str,
+        *,
+        workspace_id: Optional[str] = None,
         title: Optional[str] = None,
         description: Optional[str] = None,
         status: Optional[str] = None,
@@ -106,7 +115,7 @@ class IssueService:
         project_id: Optional[str] = None,
     ) -> Optional[Issue]:
         """Update issue fields."""
-        issue = await self.get(issue_id)
+        issue = await self.get(issue_id, workspace_id=workspace_id)
         if issue is None:
             return None
         if title is not None:
@@ -135,21 +144,27 @@ class IssueService:
         issue_id: str,
         assignee_type: str,
         assignee_id: str,
+        *,
+        workspace_id: Optional[str] = None,
     ) -> Optional[Issue]:
         """Assign an issue to a member or agent."""
         if assignee_type not in VALID_ASSIGNEE_TYPES:
             raise ValueError(f"Invalid assignee_type: {assignee_type}")
         return await self.update(
-            issue_id, assignee_type=assignee_type, assignee_id=assignee_id
+            issue_id, workspace_id=workspace_id, assignee_type=assignee_type, assignee_id=assignee_id
         )
 
-    async def transition(self, issue_id: str, new_status: str) -> Optional[Issue]:
+    async def transition(
+        self, issue_id: str, new_status: str, *, workspace_id: Optional[str] = None
+    ) -> Optional[Issue]:
         """Transition an issue to a new status."""
-        return await self.update(issue_id, status=new_status)
+        return await self.update(issue_id, workspace_id=workspace_id, status=new_status)
 
-    async def delete(self, issue_id: str) -> bool:
+    async def delete(
+        self, issue_id: str, *, workspace_id: Optional[str] = None
+    ) -> bool:
         """Delete an issue."""
-        issue = await self.get(issue_id)
+        issue = await self.get(issue_id, workspace_id=workspace_id)
         if issue is None:
             return False
         await self._session.delete(issue)

src/praisonai-platform/praisonai_platform/services/project_service.py

@@ -44,9 +44,16 @@ class ProjectService:
         await self._session.flush()
         return project
 
-    async def get(self, project_id: str) -> Optional[Project]:
-        """Get project by ID."""
-        return await self._session.get(Project, project_id)
+    async def get(
+        self, project_id: str, *, workspace_id: Optional[str] = None
+    ) -> Optional[Project]:
+        """Get project by ID, optionally scoped to a workspace."""
+        project = await self._session.get(Project, project_id)
+        if project is None:
+            return None
+        if workspace_id is not None and project.workspace_id != workspace_id:
+            return None
+        return project
 
     async def list_for_workspace(
         self,
@@ -62,6 +69,8 @@ class ProjectService:
     async def update(
         self,
         project_id: str,
+        *,
+        workspace_id: Optional[str] = None,
         title: Optional[str] = None,
         description: Optional[str] = None,
         status: Optional[str] = None,
@@ -69,7 +78,7 @@ class ProjectService:
         lead_id: Optional[str] = None,
     ) -> Optional[Project]:
         """Update project fields."""
-        project = await self.get(project_id)
+        project = await self.get(project_id, workspace_id=workspace_id)
         if project is None:
             return None
         if title is not None:
@@ -85,9 +94,11 @@ class ProjectService:
         await self._session.flush()
         return project
 
-    async def delete(self, project_id: str) -> bool:
+    async def delete(
+        self, project_id: str, *, workspace_id: Optional[str] = None
+    ) -> bool:
         """Delete a project."""
-        project = await self.get(project_id)
+        project = await self.get(project_id, workspace_id=workspace_id)
         if project is None:
             return False
         await self._session.delete(project)

src/praisonai-platform/praisonai_platform/services/workspace_service.py

@@ -10,7 +10,7 @@ from __future__ import annotations
 import re
 from typing import Optional
 
-from sqlalchemy import select
+from sqlalchemy import delete, select
 from sqlalchemy.ext.asyncio import AsyncSession
 
 from ..db.models import Member, Workspace
@@ -98,10 +98,13 @@ class WorkspaceService:
         return ws
 
     async def delete(self, workspace_id: str) -> bool:
-        """Delete a workspace."""
+        """Delete a workspace and its memberships."""
         ws = await self.get(workspace_id)
         if ws is None:
             return False
+        await self._session.execute(
+            delete(Member).where(Member.workspace_id == workspace_id)
+        )
         await self._session.delete(ws)
         await self._session.flush()
         return True

src/praisonai-platform/pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "setuptools.build_meta"
 
 [project]
 name = "praisonai-platform"
-version = "0.1.2"
+version = "0.1.4"
 description = "Platform layer for PraisonAI — workspace, auth, issues, projects"
 readme = "README.md"
 requires-python = ">=3.10"

src/praisonai-platform/tests/test_resource_idor.py

@@ -0,0 +1,20 @@
+"""Cross-workspace IDOR guards for platform resources."""
+
+from __future__ import annotations
+
+import pytest
+from fastapi import HTTPException
+
+from praisonai_platform.api.deps import ensure_resource_in_workspace
+
+
+def test_project_workspace_mismatch():
+    with pytest.raises(HTTPException) as exc:
+        ensure_resource_in_workspace("ws-other", "ws-mine", label="Project")
+    assert exc.value.status_code == 404
+
+
+def test_label_workspace_mismatch():
+    with pytest.raises(HTTPException) as exc:
+        ensure_resource_in_workspace("ws-other", "ws-mine", label="Label")
+    assert exc.value.status_code == 404

src/praisonai-platform/tests/test_service_workspace_scope.py

@@ -0,0 +1,38 @@
+"""Service-layer workspace scoping for issues and projects."""
+
+from __future__ import annotations
+
+import pytest
+
+from praisonai_platform.services.auth_service import AuthService
+from praisonai_platform.services.issue_service import IssueService
+from praisonai_platform.services.project_service import ProjectService
+from praisonai_platform.services.workspace_service import WorkspaceService
+
+
+@pytest.mark.asyncio
+async def test_issue_get_rejects_wrong_workspace(session):
+    auth = AuthService(session)
+    user, _ = await auth.register("[email protected]", "pass")
+    ws_a = await WorkspaceService(session).create("A", user.id)
+    ws_b = await WorkspaceService(session).create("B", user.id)
+    issue_svc = IssueService(session)
+    issue = await issue_svc.create(
+        workspace_id=ws_a.id,
+        title="secret",
+        creator_id=user.id,
+    )
+    assert await issue_svc.get(issue.id, workspace_id=ws_a.id) is not None
+    assert await issue_svc.get(issue.id, workspace_id=ws_b.id) is None
+
+
+@pytest.mark.asyncio
+async def test_project_delete_scoped_to_workspace(session):
+    auth = AuthService(session)
+    user, _ = await auth.register("[email protected]", "pass")
+    ws_a = await WorkspaceService(session).create("PA", user.id)
+    ws_b = await WorkspaceService(session).create("PB", user.id)
+    proj_svc = ProjectService(session)
+    project = await proj_svc.create(workspace_id=ws_a.id, title="p1")
+    assert await proj_svc.delete(project.id, workspace_id=ws_b.id) is False
+    assert await proj_svc.delete(project.id, workspace_id=ws_a.id) is True

src/praisonai-platform/tests/test_workspace_rbac.py

@@ -0,0 +1,18 @@
+"""Workspace member RBAC and cross-workspace IDOR guards."""
+
+from __future__ import annotations
+
+import pytest
+from fastapi import HTTPException
+
+from praisonai_platform.api.deps import ensure_resource_in_workspace
+
+
+def test_ensure_resource_in_workspace_rejects_mismatch():
+    with pytest.raises(HTTPException) as exc:
+        ensure_resource_in_workspace("ws-a", "ws-b", label="Issue")
+    assert exc.value.status_code == 404
+
+
+def test_ensure_resource_in_workspace_allows_match():
+    ensure_resource_in_workspace("ws-a", "ws-a", label="Issue")

src/praisonai-platform/uv.lock

@@ -1208,7 +1208,7 @@ wheels = [
 
 [[package]]
 name = "praisonai-platform"
-version = "0.1.2"
+version = "0.1.4"
 source = { editable = "." }
 dependencies = [
     { name = "aiosqlite" },

src/praisonai-rust/FEATURE_PARITY_TRACKER.json

@@ -1,6 +1,6 @@
 {
   "version": "1.5.87",
-  "lastUpdated": "2026-05-08",
+  "lastUpdated": "2026-05-20",
   "generatedBy": "praisonai._dev.parity.generator",
   "sourceOfTruth": "Python SDK (praisonaiagents)",
   "status": "PARITY_ACHIEVED",
@@ -1725,6 +1725,7 @@
       "planning": [],
       "plugins": [
         "PluginRegistry",
+        "ToolDefinition",
         "disable_plugins",
         "enable_plugins",
         "get_plugin_manager",
@@ -1749,7 +1750,6 @@
       "tools": [
         "FunctionTool",
         "Tool",
-        "ToolDefinition",
         "ToolRegistry",
         "ToolResult"
       ],

src/praisonai-ts/FEATURE_PARITY_TRACKER.json

@@ -1,6 +1,6 @@
 {
   "version": "1.5.87",
-  "lastUpdated": "2026-05-08",
+  "lastUpdated": "2026-05-20",
   "generatedBy": "praisonai._dev.parity.generator",
   "sourceOfTruth": "Python SDK (praisonaiagents)",
   "summary": {

src/praisonai-ts/PARITY.md

@@ -1,6 +1,6 @@
 # Feature Parity Tracker
 
-> **Version:** 1.5.87 | **Last Updated:** 2026-05-08
+> **Version:** 1.5.87 | **Last Updated:** 2026-05-20
 > **Source of Truth:** Python SDK (praisonaiagents)
 
 ## Summary

src/praisonai/README.md

@@ -294,8 +294,35 @@ def search(query: str) -> str:
 
 @tool
 def calculate(expression: str) -> float:
-    """Evaluate a math expression."""
-    return eval(expression)
+    """Safely evaluate a numeric arithmetic expression."""
+    import ast
+    import operator
+    
+    # Define allowed operations
+    _OPS = {
+        ast.Add: operator.add,
+        ast.Sub: operator.sub,
+        ast.Mult: operator.mul,
+        ast.Div: operator.truediv,
+        ast.Pow: operator.pow,
+        ast.USub: operator.neg,
+        ast.UAdd: operator.pos,
+    }
+    
+    def _safe_eval(node):
+        if isinstance(node, ast.Constant) and isinstance(node.value, (int, float)):
+            return node.value
+        elif isinstance(node, ast.BinOp) and type(node.op) in _OPS:
+            return _OPS[type(node.op)](_safe_eval(node.left), _safe_eval(node.right))
+        elif isinstance(node, ast.UnaryOp) and type(node.op) in _OPS:
+            return _OPS[type(node.op)](_safe_eval(node.operand))
+        else:
+            raise ValueError("Unsupported expression")
+    
+    try:
+        return _safe_eval(ast.parse(expression, mode="eval").body)
+    except (ValueError, SyntaxError, TypeError, ZeroDivisionError, OverflowError):
+        raise ValueError("Invalid arithmetic expression")
 
 agent = Agent(
     instructions="You are a helpful assistant",
@@ -304,6 +331,7 @@ agent = Agent(
 agent.start("Search for AI news and calculate 15*4")
 ```
 
+> ⚠️ **Security Note:** Never use `eval()`, `exec()`, or `subprocess` in tool functions that process LLM-generated or user-supplied input. Always validate and sanitize inputs to prevent code injection attacks.
 > 📖 [Full tools docs](https://docs.praison.ai/docs/tools/tools) — BaseTool, tool packages, 100+ built-in tools
 
 ### 5. Persistence (Databases)
@@ -330,6 +358,15 @@ pip install "praisonai[claw]"
 praisonai claw
 ```
 
+#### Required Environment Variables
+
+Copy `.env.example` to `.env` and configure the following variables:
+
+| Variable | Required | Description |
+|----------|----------|-------------|
+| `OPENAI_API_KEY` | Yes | OpenAI API key for all LLM calls |
+| `TAVILY_API_KEY` | Yes (Claw) | Tavily key for the built-in web-search tool. Get one free at https://app.tavily.com |
+
 Open **http://localhost:8082** — the dashboard comes with 13 built-in pages: Chat, Agents, Memory, Knowledge, Channels, Guardrails, Cron, and more. Add messaging channels directly from the UI.
 
 > 📖 [Full Claw docs](https://docs.praison.ai/docs/concepts/claw) — platform tokens, CLI options, Docker, and YAML agent mode

src/praisonai/praisonai.rb

@@ -3,8 +3,8 @@ class Praisonai < Formula
   
     desc "AI tools for various AI applications"
     homepage "https://github.com/MervinPraison/PraisonAI"
-    url "https://github.com/MervinPraison/PraisonAI/archive/refs/tags/v4.6.37.tar.gz"
-    sha256 `curl -sL https://github.com/MervinPraison/PraisonAI/archive/refs/tags/v4.6.37.tar.gz | shasum -a 256`.split.first
+    url "https://github.com/MervinPraison/PraisonAI/archive/refs/tags/v4.6.44.tar.gz"
+    sha256 `curl -sL https://github.com/MervinPraison/PraisonAI/archive/refs/tags/v4.6.44.tar.gz | shasum -a 256`.split.first
     license "MIT"
   
     depends_on "[email protected]"

src/praisonai/praisonai/__init__.py

@@ -36,10 +36,34 @@ __all__ = [
 _telemetry_lock = threading.Lock()
 _telemetry_initialized = False
 
+def _get_telemetry_defaults() -> dict[str, str]:
+    """Get telemetry environment defaults without mutating os.environ.
+    
+    Returns:
+        Dict of environment variable defaults for telemetry
+    """
+    import os
+    defaults = {}
+    
+    # Respect any value the user already set
+    if "OTEL_SDK_DISABLED" not in os.environ:
+        langfuse_configured = bool(
+            os.getenv("LANGFUSE_PUBLIC_KEY")
+            or os.path.exists(os.path.expanduser("~/.praisonai/langfuse.env"))
+        )
+        defaults["OTEL_SDK_DISABLED"] = "false" if langfuse_configured else "true"
+    
+    if "EC_TELEMETRY" not in os.environ:
+        defaults["EC_TELEMETRY"] = "false"
+    
+    return defaults
+
+
 def _ensure_telemetry_defaults() -> None:
     """Apply telemetry env defaults exactly once, on first observability use.
     
     Thread-safe implementation using double-checked locking pattern.
+    DEPRECATED: Use _apply_telemetry_defaults() with explicit config instead.
     """
     global _telemetry_initialized
     if _telemetry_initialized:  # fast path, OK without lock
@@ -47,18 +71,21 @@ def _ensure_telemetry_defaults() -> None:
     with _telemetry_lock:
         if _telemetry_initialized:
             return
-        import os
-        # Respect any value the user already set
-        if "OTEL_SDK_DISABLED" not in os.environ:
-            langfuse_configured = bool(
-                os.getenv("LANGFUSE_PUBLIC_KEY")
-                or os.path.exists(os.path.expanduser("~/.praisonai/langfuse.env"))
-            )
-            os.environ["OTEL_SDK_DISABLED"] = "false" if langfuse_configured else "true"
-        os.environ.setdefault("EC_TELEMETRY", "false")  # respect user overrides
+        _apply_telemetry_defaults(_get_telemetry_defaults())
         _telemetry_initialized = True
 
 
+def _apply_telemetry_defaults(env_vars: dict[str, str]) -> None:
+    """Apply telemetry defaults to os.environ.
+    
+    Args:
+        env_vars: Dict of environment variables to set
+    """
+    import os
+    for key, value in env_vars.items():
+        os.environ.setdefault(key, value)
+
+
 # Lazy loading for heavy imports
 def __getattr__(name):
     """Lazy load heavy modules to improve import time."""

src/praisonai/praisonai/_framework_availability.py

@@ -0,0 +1,53 @@
+"""Single source of truth for optional-framework availability."""
+import importlib.util
+import importlib.metadata
+import threading
+from typing import Callable
+
+_cache: dict[str, bool] = {}
+_lock = threading.Lock()
+
+def _ag2_probe() -> bool:
+    # AG2 ships under the `autogen` namespace; presence of the dist is
+    # the only reliable signal.
+    try:
+        importlib.metadata.distribution('ag2')
+    except importlib.metadata.PackageNotFoundError:
+        return False
+    return importlib.util.find_spec("autogen") is not None
+
+def _autogen_v4_probe() -> bool:
+    return (importlib.util.find_spec("autogen_agentchat") is not None
+            and importlib.util.find_spec("autogen_ext") is not None)
+
+_PROBES: dict[str, Callable[[], bool]] = {
+    "crewai":            lambda: importlib.util.find_spec("crewai") is not None,
+    "autogen":           lambda: importlib.util.find_spec("autogen") is not None,
+    "autogen_v4":        _autogen_v4_probe,
+    "ag2":               _ag2_probe,
+    "praisonaiagents":   lambda: importlib.util.find_spec("praisonaiagents") is not None,
+    "praisonai_tools":   lambda: importlib.util.find_spec("praisonai_tools") is not None,
+    "agentops":          lambda: importlib.util.find_spec("agentops") is not None,
+    "litellm":           lambda: importlib.util.find_spec("litellm") is not None,
+    "openai":            lambda: importlib.util.find_spec("openai") is not None,
+}
+
+def is_available(name: str) -> bool:
+    if name not in _PROBES:
+        raise ValueError(f"unknown framework name: {name!r}")
+    cached = _cache.get(name)
+    if cached is not None:
+        return cached
+    with _lock:
+        cached = _cache.get(name)
+        if cached is None:
+            cached = bool(_PROBES[name]())
+            _cache[name] = cached
+        return cached
+
+def invalidate(name: str | None = None) -> None:
+    with _lock:
+        if name is None:
+            _cache.clear()
+        else:
+            _cache.pop(name, None)

src/praisonai/praisonai/_registry.py

@@ -0,0 +1,228 @@
+"""Generic plugin registry implementation for PraisonAI.
+
+This module provides a unified registry pattern to replace the divergent singleton
+implementations across the codebase. It supports both built-in and entry-point
+based plugin discovery with thread-safe registration and dependency injection.
+"""
+
+from __future__ import annotations
+
+import threading
+from importlib.metadata import entry_points
+from typing import Callable, Dict, Generic, Optional, Type, TypeVar
+import logging
+
+T = TypeVar("T")
+logger = logging.getLogger(__name__)
+
+
+class PluginRegistry(Generic[T]):
+    """Generic plugin registry: builtins + entry points + runtime register().
+    
+    This replaces the singleton pattern with dependency injection while maintaining
+    the same functionality. Supports:
+    - Built-in plugins with lazy loading
+    - Entry points discovery
+    - Runtime registration
+    - Alias support for alternative names
+    - Thread-safe operations
+    """
+
+    def __init__(
+        self, 
+        *,
+        entry_point_group: str,
+        builtins: Optional[Dict[str, Callable[[], Type[T]]]] = None
+    ) -> None:
+        """Initialize the registry.
+        
+        Args:
+            entry_point_group: Entry points group name for plugin discovery
+            builtins: Dict of name -> loader function for built-in plugins
+        """
+        self._entry_point_group = entry_point_group
+        self._items: Dict[str, Type[T]] = {}
+        self._aliases: Dict[str, str] = {}  # alias -> canonical name
+        self._lock = threading.Lock()
+        
+        # Load built-in plugins with error handling
+        if builtins:
+            for name, loader in builtins.items():
+                try:
+                    self._items[name] = loader()
+                except ImportError:
+                    # Built-in plugin dependencies not available, skip
+                    pass
+                except Exception:
+                    # Log other errors but don't crash initialization
+                    logger.warning("Failed to load built-in plugin %r", name, exc_info=True)
+        
+        self._load_entry_points()
+
+    def _load_entry_points(self) -> None:
+        """Load plugins from entry points."""
+        try:
+            for ep in entry_points(group=self._entry_point_group):
+                try:
+                    plugin_class = ep.load()
+                    with self._lock:
+                        self._items[ep.name] = plugin_class
+                except Exception:
+                    # Do not break plugin dispatch because one plugin is broken
+                    logger.warning(
+                        "Failed to load plugin %r from entry point",
+                        ep.name,
+                        exc_info=True,
+                    )
+        except Exception:
+            # entry_points() might not be available in older Python versions
+            logger.debug("Entry points not available for group %s", self._entry_point_group)
+
+    def register(
+        self, 
+        name: str, 
+        cls: Type[T], 
+        *, 
+        aliases: Optional[list[str]] = None
+    ) -> None:
+        """Register a plugin at runtime.
+        
+        Args:
+            name: Plugin name
+            cls: Plugin class
+            aliases: Optional list of alias names for this plugin
+        """
+        with self._lock:
+            canonical_name = name.lower()
+            self._items[canonical_name] = cls
+            
+            # Register aliases
+            if aliases:
+                for alias in aliases:
+                    normalized_alias = alias.lower()
+                    self._aliases[normalized_alias] = canonical_name
+
+    def unregister(self, name: str) -> bool:
+        """Unregister a plugin.
+        
+        Args:
+            name: Plugin name
+            
+        Returns:
+            True if plugin was found and removed, False otherwise
+        """
+        with self._lock:
+            normalized_name = name.lower()
+            
+            # Check if it's an alias
+            if normalized_name in self._aliases:
+                del self._aliases[normalized_name]
+                return True
+            
+            # Check if it's a canonical name
+            if normalized_name in self._items:
+                # Remove all aliases pointing to this plugin
+                aliases_to_remove = [
+                    alias for alias, canonical in self._aliases.items()
+                    if canonical == normalized_name
+                ]
+                for alias in aliases_to_remove:
+                    del self._aliases[alias]
+                
+                del self._items[normalized_name]
+                return True
+            
+            return False
+
+    def resolve(self, name: str) -> Type[T]:
+        """Resolve a plugin name to its class.
+        
+        Args:
+            name: Plugin name
+            
+        Returns:
+            Plugin class
+            
+        Raises:
+            ValueError: If plugin is not found
+        """
+        with self._lock:
+            normalized_name = name.lower()
+            
+            # Resolve alias to canonical name
+            canonical_name = self._aliases.get(normalized_name, normalized_name)
+            cls = self._items.get(canonical_name)
+            
+            # Capture available plugins snapshot while holding lock
+            # to avoid race condition between check and error message
+            if cls is None:
+                available_plugins = sorted(self._items.keys())
+                available_aliases = sorted(self._aliases.keys())
+                available_snapshot = available_plugins + available_aliases
+            else:
+                available_snapshot = None
+        
+        if cls is None:
+            raise ValueError(
+                f"Unknown {self._entry_point_group} plugin: {name!r}. "
+                f"Available: {available_snapshot}"
+            )
+        return cls
+
+    def create(self, name: str, *args, **kwargs) -> T:
+        """Create an instance of the specified plugin.
+        
+        Args:
+            name: Plugin name
+            *args, **kwargs: Arguments to pass to plugin constructor
+            
+        Returns:
+            Plugin instance
+            
+        Raises:
+            ValueError: If plugin is not found
+        """
+        cls = self.resolve(name)
+        return cls(*args, **kwargs)
+
+    def list_names(self) -> list[str]:
+        """List all registered plugin names.
+        
+        Returns:
+            Sorted list of plugin names
+        """
+        with self._lock:
+            return sorted(self._items.keys())
+
+    def is_available(self, name: str) -> bool:
+        """Check if a plugin is available.
+        
+        Args:
+            name: Plugin name
+            
+        Returns:
+            True if plugin exists and can be created
+        """
+        try:
+            self.resolve(name)
+            return True
+        except ValueError:
+            return False
+
+    def list_aliases(self) -> Dict[str, str]:
+        """List all aliases and their target plugins.
+        
+        Returns:
+            Dict mapping alias -> canonical name
+        """
+        with self._lock:
+            return dict(self._aliases)
+
+    def list_all_names(self) -> list[str]:
+        """List all names including aliases.
+        
+        Returns:
+            Sorted list of all registered names and aliases
+        """
+        with self._lock:
+            return sorted(list(self._items.keys()) + list(self._aliases.keys()))

src/praisonai/praisonai/agents_generator.py

@@ -19,16 +19,12 @@ import difflib
 
 # Import new architecture components
 from .framework_adapters.base import FrameworkAdapter
-from .framework_adapters.registry import FrameworkAdapterRegistry
+from .framework_adapters.registry import FrameworkAdapterRegistry, get_default_registry
 from .tool_registry import ToolRegistry
 
 # Import availability flags
-try:
-    from .inbuilt_tools import PRAISONAI_TOOLS_AVAILABLE, CREWAI_AVAILABLE, AUTOGEN_AVAILABLE
-except ImportError:
-    PRAISONAI_TOOLS_AVAILABLE = False
-    CREWAI_AVAILABLE = False
-    AUTOGEN_AVAILABLE = False
+# Compatibility imports - now handled by centralized detection
+# (inbuilt_tools still defines these but they're read-only compatibility)
 
 # Import BaseTool for tools handling
 BaseTool = None
@@ -40,15 +36,14 @@ except ImportError:
     except ImportError:
         pass
 
-# Check for additional framework availability
-AG2_AVAILABLE = False
-PRAISONAI_AVAILABLE = False
-try:
-    import importlib.util
-    AG2_AVAILABLE = importlib.util.find_spec("ag2") is not None
-    PRAISONAI_AVAILABLE = importlib.util.find_spec("praisonaiagents") is not None
-except ImportError:
-    pass
+# Check for additional framework availability using centralized detection
+from ._framework_availability import is_available
+PRAISONAI_TOOLS_AVAILABLE = is_available("praisonai_tools")
+CREWAI_AVAILABLE          = is_available("crewai")
+AUTOGEN_AVAILABLE         = is_available("autogen")
+AG2_AVAILABLE             = is_available("ag2")
+PRAISONAI_AVAILABLE       = is_available("praisonaiagents")
+AGENTOPS_AVAILABLE        = is_available("agentops")
 
 # Framework adapter registry - now uses proper registry pattern
 # This replaces the hardcoded FRAMEWORK_ADAPTERS dict
@@ -179,7 +174,7 @@ def _resolve_yaml_cli_backend(cli_backend_config, logger):
 
 
 class AgentsGenerator:
-    def __init__(self, agent_file, framework, config_list, log_level=None, agent_callback=None, task_callback=None, agent_yaml=None, tools=None, cli_config=None):
+    def __init__(self, agent_file, framework, config_list, log_level=None, agent_callback=None, task_callback=None, agent_yaml=None, tools=None, cli_config=None, adapter_registry=None):
         """
         Initialize the AgentsGenerator object.
 
@@ -193,6 +188,7 @@ class AgentsGenerator:
             agent_yaml (str, optional): The content of the YAML file. Defaults to None.
             tools (dict, optional): A dictionary containing the tools to be used for the agents. Defaults to None.
             cli_config (dict, optional): CLI configuration to override YAML settings. Defaults to None.
+            adapter_registry (FrameworkAdapterRegistry, optional): Registry for framework adapters. Defaults to process default.
 
         Attributes:
             agent_file (str): The path to the agent file.
@@ -233,6 +229,10 @@ class AgentsGenerator:
         self.tool_registry = ToolRegistry()
         self.tool_registry.register_builtin_autogen_adapters()
         
+        # DI-friendly: tests/multi-tenant runtimes pass their own registry;
+        # CLI users get the process default.
+        self._adapter_registry = adapter_registry or get_default_registry()
+        
         # Get framework adapter (availability already validated at CLI entry)
         self.framework_adapter = self._get_framework_adapter(framework)
 
@@ -249,8 +249,7 @@ class AgentsGenerator:
         Raises:
             ValueError: If framework is not supported
         """
-        adapter_registry = FrameworkAdapterRegistry.get_instance()
-        return adapter_registry.create(framework)
+        return self._adapter_registry.create(framework)
 
     def _merge_cli_config(self, config, cli_config):
         """
@@ -375,20 +374,19 @@ class AgentsGenerator:
 
     def load_tools_from_module(self, module_path):
         """
-        Loads tools from a specified module path.
+        Load function tools from a user-supplied module (gated by PRAISONAI_ALLOW_LOCAL_TOOLS).
 
         Parameters:
             module_path (str): The path to the module containing the tools.
 
         Returns:
             dict: A dictionary containing the names of the tools as keys and the corresponding functions or objects as values.
-
-        Raises:
-            FileNotFoundError: If the specified module path does not exist.
+                  Returns an empty dict if the module cannot be loaded (path missing, loading blocked by PRAISONAI_ALLOW_LOCAL_TOOLS, or any other load error).
         """
-        spec = importlib.util.spec_from_file_location("tools_module", module_path)
-        module = importlib.util.module_from_spec(spec)
-        spec.loader.exec_module(module)
+        from ._safe_loader import load_user_module
+        module = load_user_module(module_path, name="tools_module")
+        if module is None:
+            return {}
         return {name: obj for name, obj in inspect.getmembers(module, self.is_function_or_decorated)}
     
     def _extract_tool_classes(self, module):
@@ -411,21 +409,13 @@ class AgentsGenerator:
     
     def load_tools_from_module_class(self, module_path):
         """
-        Loads tools from a specified module path containing classes that inherit from BaseTool 
-        or are part of langchain_community.tools package.
+        Load BaseTool / langchain tool classes from a user-supplied module (gated by PRAISONAI_ALLOW_LOCAL_TOOLS).
         """
-        spec = importlib.util.spec_from_file_location("tools_module", module_path)
-        module = importlib.util.module_from_spec(spec)
-        try:
-            spec.loader.exec_module(module)
-            return {name: obj() for name, obj in inspect.getmembers(module, 
-                lambda x: inspect.isclass(x) and (
-                    x.__module__.startswith('langchain_community.tools') or 
-                    (PRAISONAI_TOOLS_AVAILABLE and BaseTool and issubclass(x, BaseTool))
-                ) and x is not BaseTool)}
-        except ImportError as e:
-            self.logger.warning(f"Error loading tools from {module_path}: {e}")
+        from ._safe_loader import load_user_module
+        module = load_user_module(module_path, name="tools_module")
+        if module is None:
             return {}
+        return self._extract_tool_classes(module)
 
     def load_tools_from_package(self, package_path):
         """
@@ -451,36 +441,6 @@ class AgentsGenerator:
                     tools_dict[name] = obj
         return tools_dict
 
-    def load_tools_from_tools_py(self):
-        """
-        Imports and returns all contents from tools.py file.
-        Uses the tool registry instead of global namespace pollution.
-
-        Returns:
-            list: A list of callable functions with proper formatting
-        """
-        tools_list = []
-        try:
-            # Try to import tools.py from current directory using safe loading
-            from ._safe_loader import load_user_module
-            module = load_user_module("tools.py", name="tools")
-            if module is None:
-                self.logger.debug("tools.py not found or local tools loading disabled")
-                return tools_list
-
-            # Register functions in the tool registry instead of globals()
-            registered_tools = self.tool_registry.register_from_module(module)
-            tools_list = [self.tool_registry.get_function(name) for name in registered_tools]
-
-            self.logger.debug(f"Loaded {len(tools_list)} tool functions from tools.py")
-            self.logger.debug(f"Registered tools: {registered_tools}")
-            
-        except FileNotFoundError:
-            self.logger.debug("tools.py not found in current directory")
-        except Exception as e:
-            self.logger.warning(f"Error loading tools from tools.py: {e}")
-            
-        return tools_list
 
     def generate_crew_and_kickoff(self):
         """
@@ -590,12 +550,10 @@ class AgentsGenerator:
         tools_py_path = os.path.join(root_directory, 'tools.py')
         tools_dir_path = Path(root_directory) / 'tools'
         
+        # Use consolidated ToolResolver for tools.py loading
+        tools_dict.update(self.tool_resolver.get_local_tool_classes())
         if os.path.isfile(tools_py_path):
-            from ._safe_loader import load_user_module
-            module = load_user_module(tools_py_path, name="tools_module")
-            if module is not None:
-                tools_dict.update(self._extract_tool_classes(module))
-                self.logger.debug("tools.py exists in the root directory. Loading tools.py and skipping tools folder.")
+            self.logger.debug("tools.py exists in the root directory. Loading tools.py and skipping tools folder.")
         elif tools_dir_path.is_dir():
             from ._safe_loader import load_user_module
             for py_file in tools_dir_path.glob("*.py"):
@@ -784,7 +742,11 @@ class AgentsGenerator:
         result = "### Output ###\n" + response[-1].summary if hasattr(response[-1], 'summary') else ""
         
         if AGENTOPS_AVAILABLE:
-            agentops.end_session("Success")
+            import agentops
+            try:
+                agentops.end_session("Success")
+            except Exception as e:  # noqa: BLE001 -- agentops errors must not crash the caller
+                self.logger.warning(f"agentops.end_session failed: {e}")
             
         return result
 
@@ -1171,7 +1133,11 @@ class AgentsGenerator:
         result = f"### Task Output ###\n{response}"
         
         if AGENTOPS_AVAILABLE:
-            agentops.end_session("Success")
+            import agentops
+            try:
+                agentops.end_session("Success")
+            except Exception as e:  # noqa: BLE001 -- agentops errors must not crash the caller
+                self.logger.warning(f"agentops.end_session failed: {e}")
             
         return result
 
@@ -1190,8 +1156,8 @@ class AgentsGenerator:
 
         # Use existing tool resolver instance
         
-        # Load tools from local tools.py (backward compat)
-        tools_list = self.load_tools_from_tools_py()
+        # Load tools from local tools.py (backward compat) - use consolidated ToolResolver
+        tools_list = self.tool_resolver.get_local_callables()
         self.logger.debug(f"Loaded tools from tools.py: {tools_list}")
 
         # Initialize InteractiveRuntime for ACP/LSP if enabled globally
@@ -1199,9 +1165,9 @@ class AgentsGenerator:
         acp_enabled = global_config.get('acp', False)
         lsp_enabled = global_config.get('lsp', False)
         interactive_runtime = None
-        interactive_loop = None
         
         if acp_enabled or lsp_enabled:
+            runtime_started = False
             try:
                 import asyncio
                 from praisonai.cli.features.interactive_runtime import InteractiveRuntime, RuntimeConfig
@@ -1217,27 +1183,34 @@ class AgentsGenerator:
                 interactive_runtime = InteractiveRuntime(runtime_config)
                 self.logger.info(f"Starting InteractiveRuntime (ACP: {acp_enabled}, LSP: {lsp_enabled})")
                 
-                # Create a scoped event loop instead of modifying process globals
-                interactive_loop = asyncio.new_event_loop()
-                try:
-                    interactive_loop.run_until_complete(interactive_runtime.start())
-                    
-                    centric_tools = create_agent_centric_tools(interactive_runtime)
-                    self.logger.info(f"Loaded {len(centric_tools)} InteractiveRuntime tools")
-                    tools_list.extend(centric_tools)
-                    
-                finally:
-                    try:
-                        interactive_loop.run_until_complete(interactive_runtime.stop())
-                    except Exception as stop_error:
-                        self.logger.warning(f"Error stopping InteractiveRuntime: {stop_error}")
-                    finally:
-                        interactive_loop.close()
+                # Runs on the persistent background loop; safe from sync and async callers.
+                # run_sync raises RuntimeError early if called from inside a running loop
+                # so the bug is loud instead of a deadlock.
+                from ._async_bridge import run_sync
+                run_sync(interactive_runtime.start())
+                runtime_started = True
+                
+                centric_tools = create_agent_centric_tools(interactive_runtime)
+                self.logger.info(f"Loaded {len(centric_tools)} InteractiveRuntime tools")
+                tools_list.extend(centric_tools)
                 
             except ImportError as e:
                 self.logger.warning(f"Failed to load InteractiveRuntime components: {e}")
+                interactive_runtime = None
+            except RuntimeError:
+                # Don't swallow RuntimeError from run_sync - preserve fail-fast semantics
+                raise
             except Exception as e:
+                if runtime_started and interactive_runtime is not None:
+                    try:
+                        from ._async_bridge import run_sync
+                        run_sync(interactive_runtime.stop())
+                    except Exception as stop_error:
+                        self.logger.error(
+                            f"Error stopping partially started InteractiveRuntime: {stop_error}"
+                        )
                 self.logger.error(f"Error starting InteractiveRuntime: {e}")
+                interactive_runtime = None
 
         # Create agents from config
         for role, details in config['roles'].items():
@@ -1471,14 +1444,19 @@ class AgentsGenerator:
             self.logger.debug(f"Result: {response}")
             result = response if response else ""
         finally:
-            if interactive_runtime and interactive_loop:
+            if interactive_runtime:
                 try:
                     self.logger.info("Stopping InteractiveRuntime...")
-                    interactive_loop.run_until_complete(interactive_runtime.stop())
+                    from ._async_bridge import run_sync
+                    run_sync(interactive_runtime.stop())
                 except Exception as e:
                     self.logger.error(f"Error stopping InteractiveRuntime: {e}")
         
         if AGENTOPS_AVAILABLE:
-            agentops.end_session("Success")
+            import agentops
+            try:
+                agentops.end_session("Success")
+            except Exception as e:  # noqa: BLE001 -- agentops errors must not crash the caller
+                self.logger.warning(f"agentops.end_session failed: {e}")
             
         return result

src/praisonai/praisonai/api/agent_invoke.py

@@ -29,14 +29,26 @@ logger = logging.getLogger(__name__)
 # Authentication
 import os
 CALL_SERVER_TOKEN = os.getenv('CALL_SERVER_TOKEN')
+_CALL_AUTH_DISABLED = os.getenv('PRAISONAI_CALL_AUTH', '').lower() == 'disabled'
+
 
 async def verify_token(
     request: Request, 
     authorization: Optional[str] = Header(None)
 ) -> None:
     """Verify API token for authentication."""
-    if not FASTAPI_AVAILABLE or not CALL_SERVER_TOKEN:
-        return  # No authentication if FastAPI unavailable or no token set
+    if not FASTAPI_AVAILABLE:
+        return
+    if _CALL_AUTH_DISABLED:
+        return
+    if not CALL_SERVER_TOKEN:
+        raise HTTPException(
+            status_code=503,
+            detail=(
+                "CALL_SERVER_TOKEN is not configured. Set CALL_SERVER_TOKEN or "
+                "PRAISONAI_CALL_AUTH=disabled to run without authentication."
+            ),
+        )
         
     token = None