vxcontrol/pentagi
1
0
Fork 0
mirror of https://github.com/vxcontrol/pentagi.git synced 2026-07-20 23:57:11 -04:00
Code Issues 45 Packages Projects Releases 9 Wiki Activity

9 Releases 9 Tags

RSS Feed
  • v2.1.0 a112db206b
    Compare

    Stable

    yindo released this 2026-05-29 09:16:53 -04:00 | 5 commits to main since this release

    📅 Originally published on GitHub: Fri, 29 May 2026 18:25:21 GMT
    🏷️ Git tag created: Fri, 29 May 2026 13:16:53 GMT

    PentAGI 2.1 — File Management, Knowledge Base, ToolCall Observability, and Assistant Flow Control

    This release adds a complete file management layer (user resource libraries and flow workspace files with container sync), a first-class Knowledge Base with semantic search and anonymization, real-time ToolCall logging, and assistant tools to monitor and steer running flows. It also refreshes model configurations across OpenAI, Anthropic, Gemini, DeepSeek, Qwen, Kimi, and GLM, a broad frontend modernization pass (React 19, Apollo Client v4, Vite 8), and a wide range of stability and security fixes.

    Discord⠀Telegram

    Major Features

    User Resources & Flow Files

    A new file management layer lets users bring their own files into PentAGI and share them with agents.

    • User Resources — a persistent, per-user file library with MD5-deduplicated storage and a virtual path filesystem. Full REST and GraphQL CRUD (upload, mkdir, move, copy, delete, download), real-time subscriptions, and atomic multi-source / multi-path batch operations.
    • Flow files — per-flow workspace files that sync into worker containers at /work/uploads and /work/resources. Files can be pulled back out of a running container and promoted into the user library. Attached files are injected into agent system prompts as a structured <task_files> block, so the assistant is aware of what the user provided.
    • File Manager UI — a reusable tree component with multi-select, keyboard navigation, drag-and-drop, sortable columns, bulk actions, and an overwrite workflow. Used by both the new /resources page and the flow Files tab.
    • Limits and hardening — enforced on both ends: 300 MB per file, 1000 files per request, 2 GB total, 255-byte names. Upload paths are protected against directory traversal and symlink escapes.

    Knowledge Base Management

    The pgvector memory store is now a first-class, user-manageable resource rather than agent-only auto-storage.

    • GraphQL/REST CRUD plus semantic search over the knowledge store, with admin/user scoping, per-user document ownership, re-embedding on update, and real-time subscriptions.
    • A new /knowledges interface with list and detail pages, a TipTap markdown editor, partial updates, inline rename/delete, and a collapsible semantic-search input (hotkey-accessible).
    • Text anonymization — a new anonymizeText service (GraphQL/REST) to scrub sensitive data, surfaced directly in the knowledge editor.
    • The vector search engine was rewritten onto direct, parameterized SQL queries — fixing a bug where document IDs were dropped and removing unsafe SQL string interpolation.

    ToolCall Observability

    Individual agent tool calls are now logged through a dedicated provider and exposed via GraphQL queries and subscriptions plus a REST API, giving real-time, inspectable visibility into every command, search, and action an agent performs during a flow.

    Assistant Flow Management

    The interactive assistant can now observe and steer active flows without leaving the chat. New tools let the assistant read flow status (with multiple detail levels), stop a flow, submit input to a waiting flow, patch subtasks, and block until a task completes. A summarizer cache avoids redundant LLM calls when reporting flow state.

    Unified Agent Language Policy

    A single, consistent language policy now spans all agent prompts: vector-store and search-engine queries are forced to English for retrieval consistency, while user-facing messages follow the engagement language. Template variables and tool access were aligned with each agent's actual runtime tool set.

    New Capabilities

    Updated Provider Configurations & Models

    Built-in model configurations were refreshed across multiple providers — OpenAI, Anthropic, Gemini, DeepSeek, Qwen, Kimi, and GLM — to match the current model landscape, with updated model lists, pricing, and context windows.

    • Per-role thinking control — reasoning models such as DeepSeek and Qwen now toggle thinking mode per agent role, so utility roles run without thinking (and honor their sampling parameters) while reasoning, tool-use, and security-analysis roles keep it enabled.
    • New reference configurations — vLLM Qwen 3.6 (thinking and non-thinking, including 35B FP8 variants) bundled in the Docker image, and an Azure OpenAI example.
    • Ollama now surfaces a clear, actionable error when the selected model does not support tool/function calling, instead of a deeply nested stack trace.

    Per-Model Analytics

    A new query surfaces token usage broken down by model and agent type within a single flow, integrated into the flow dashboard.

    Configuration

    • TERMINAL_TOOL_TIMEOUT — configurable terminal command timeout (default raised to 1200 seconds, with clamping for out-of-range values).
    • PostgreSQL connection pooling — shared pools for sqlc, GORM, and pgvector with new tunables DB_MAX_OPEN_CONNS, DB_MAX_IDLE_CONNS, and DB_VECTOR_MAX_CONNS.
    • EMBEDDING_MAX_TEXT_BYTES — caps the text size sent to the embedding model.
    • The Settings API now exposes version and isDevelopMode.

    Frontend Modernization

    A broad pass touched nearly every list and detail page.

    • Unified list tables — URL-synced filtering, pagination, sorting, and column visibility, with multi-column search ("Search in" column picker) and contextual empty states.
    • Detail navigation — Prev/Position/Next navigation between sibling records, with an in-sheet searchable list.
    • Inline actions — rename, finish, and delete directly from flow, template, and knowledge headers and list rows.
    • Per-route document titles — browser tabs now reflect the actual page (including live flow titles), driven by a centralized title registry.
    • Mobile UX — responsive headers that collapse to icon-only buttons, a unified flow attachment/template picker, and a compact dashboard period switcher.
    • Performance — dashboard period-switch interaction latency reduced from 434 ms to 134 ms, the PDF renderer is lazy-loaded (cutting the report route's initial JS by ~1.5 MB), the knowledge provider is scoped to its own routes (avoiding a ~2.1 MB payload on every page), filtering is debounced, and rename/favorite actions update optimistically for instant feedback.
    • Platform upgrades — React 19, Apollo Client v4, Vite 8 (Rolldown), TypeScript 6, Zod v4, the graphql-codegen v6/v7 toolchain, and the shadcn new-york-v4 component style. The frontend test suite grew from 475 to 541 tests.
    • Accessibility — aria-labels across icon-only buttons, form-field id/name fixes, and Radix dialog compliance.

    Bug Fixes & Reliability

    Flow & Agent Execution

    • Task cancellation — subtask generation now runs under a cancellable context, so cancelling a task no longer reports a false success.
    • Custom prompts — user prompt overrides saved in Settings → Prompts are now actually applied to new assistant and flow sessions (they were silently using the defaults).
    • Malformed tool-call JSON — truncated or invalid LLM arguments now fall back to an empty object instead of triggering LiteLLM 400 errors and infinite retry loops; literal control characters in arguments are sanitized before storage.
    • Subscription backpressure — events are dropped for slow or disconnected subscribers after a timeout, preventing goroutine accumulation.
    • Deadlock fixes — resolved a deadlock in the log worker and a nil-channel deadlock when finishing an assistant session.
    • Browser tool — small/empty page content now returns a warning rather than an error, binary URLs are reported clearly, and a failed screenshot no longer discards successfully fetched page content.

    Knowledge & Data

    • Vector search safety hardened (parameterized queries, memory documents excluded at the SQL level).
    • Fixed recursive resource retrieval over GraphQL, and resource move/copy responses now return the correct entries for client cache consistency.

    Frontend

    • Fixed a production crash on flow detail pages caused by the minifier stripping function names from document-title components.
    • Eliminated several table state races (filter clearing, pagination URL loops, batched URL updates) and a GraphQL codegen issue that emitted duplicate types and broke the dev server.
    • API token names are no longer lost when a subscription refetches the table mid-edit; the default button type no longer triggers accidental form submits.

    Security

    • Flow file uploads are hardened against path traversal and symlink escapes, with size and count limits enforced on both the backend and the frontend.
    • Knowledge vector search uses parameterized queries, removing prior string-interpolated SQL.
    • New endpoints enforce user/admin privilege scoping, with dedicated privileges (anonymize.call, toolcall access) added via migration.
    • Text anonymization is available to scrub sensitive data from stored knowledge.

    Documentation

    Extensive user-facing documentation was added, including a first-use guide, a pentesting prompt methodology guide, memory lifecycle across flows, capability boundaries, OAuth callback setup, a Docker mirror guide for restricted networks, OSINT integration scenarios, the flow Files tab, DeepSeek V4 migration and pricing, and a clarification that Vertex AI is reachable today only via an OpenAI-compatible gateway. Two design RFCs — flow concurrency with completion webhooks, and MCP client integration — were added under examples/proposals/ as design proposals with no runtime code yet.

    Upgrade Notes

    • DeepSeek: deployments using the legacy deepseek-chat / deepseek-reasoner model names should migrate to deepseek-v4-flash / deepseek-v4-pro before the upstream deprecation on 2026-07-24.
    • Database: connection-pool settings were consolidated to DB_MAX_OPEN_CONNS, DB_MAX_IDLE_CONNS, and DB_VECTOR_MAX_CONNS — verify against .env.example.
    • Terminal timeout: TERMINAL_TOOL_TIMEOUT default raised from 600 to 1200 seconds; review if a lower value was intentional.
    • Frontend development now requires pnpm (previously npm).
    • Database migrations apply automatically at startup. After pulling, rebuild and restart: docker compose build && docker compose up -d.

    Contributors

    Core Team

    • @asdek (Dmitry Nagibin) — User resources & flow files backend, knowledge base API and vector search, ToolCall logging, anonymizer, assistant flow management tools, agent language policy, provider model updates, database connection pooling, and flow reliability fixes
    • @sirozha (Sergey Kozyrenko) — File Manager component, resources/knowledges/flow-files UI, unified list tables and multi-column search, detail navigation, document titles, mobile UX, frontend platform upgrade (React 19 / Apollo v4 / Vite 8), and performance & accessibility work

    External Contributors

    • @mason5052 — Custom prompts fix, Ollama tool-support error clarity, DeepSeek V4 migration, flow file upload hardening, and extensive documentation and design RFCs (flow concurrency, MCP client integration, evidence chain)
    • @Kairos-T — Documentation mermaid syntax fix (PR#259)

    Full Changelog: https://github.com/vxcontrol/pentagi/compare/v2.0.0...v2.1.0

    Downloads
  • v2.0.0 87ac00f195
    Compare

    Stable

    yindo released this 2026-04-11 08:20:52 -04:00 | 314 commits to main since this release

    📅 Originally published on GitHub: Sat, 11 Apr 2026 22:10:02 GMT
    🏷️ Git tag created: Sat, 11 Apr 2026 12:20:52 GMT

    PentAGI 2.0 — Broader Model Support, Analytics, Runtime Flexibility, and Agent Guardrails

    This release expands the LLM provider ecosystem with four new providers, introduces a full analytics dashboard, enables runtime provider switching mid-flow, adds Docker host network mode for OOB attack scenarios, and ships a comprehensive set of stability and reliability improvements. It also includes significant test coverage expansion across the backend codebase.

    Discord⠀Telegram

    Major Features

    Four New LLM Providers: DeepSeek, GLM, Kimi, Qwen

    Native support for four providers from the Chinese LLM ecosystem — DeepSeek, GLM (Zhipu AI), Kimi (Moonshot AI), and Qwen (Alibaba Cloud). Each is available through the standard provider configuration interface with API key and server URL environment variables (DEEPSEEK_API_KEY, GLM_API_KEY, KIMI_API_KEY, QWEN_API_KEY). All four providers are registered in the GraphQL schema and settings UI, and come with pre-configured model lists and pricing information. Unit test coverage for all four providers is included (~71% coverage per provider).

    Ollama Cloud Support

    In addition to local Ollama deployments, PentAGI now ships a pre-built ollama-cloud.provider.yml configuration with 7 cloud-hosted models assigned per agent type: nemotron-3-super, qwen3-coder-next, glm-5, minimax-m2.7, qwen3.5:397b, and devstral-2:123b. Both Free Tier and Paid Tier setup options are documented. The configuration is bundled in the Docker image at /opt/pentagi/conf/ollama-cloud.provider.yml.

    Analytics Dashboards

    A new analytics dashboard surfaces usage statistics and cost data collected from the REST API analytics endpoints introduced in v1.2.0. The dashboard shows:

    • Token usage and cost breakdown per flow and per agent type (primary, pentester, coder, installer, searcher, adviser, etc.)
    • Cache hit rates and cache read/write cost separation for Anthropic and Gemini providers
    • Tool call frequency and execution time metrics per flow and subtask
    • Per-model cost detail, useful when running multiple provider configurations simultaneously

    Runtime Provider Switching

    It is now possible to switch the active LLM provider for a running flow without restarting the application. To switch providers: pause the flow using the stop button, navigate to provider settings to change the active provider or update configuration, then resume the flow with a message directing the agent. The backend applies conditional chain normalization to preserve the reasoning cache when the provider is unchanged, and converts tool call IDs when switching providers. The modelProvider parameter has been added to the relevant GraphQL mutations to support this workflow.

    Agent Supervision System (Beta)

    Two optional supervision mechanisms are now available and disabled by default:

    • AGENT_PLANNING_STEP_ENABLED=true — enables a planning step before each specialist agent starts work, where a planner generates a 3–7 step execution plan to scope the subtask and prevent drift.
    • EXECUTION_MONITOR_ENABLED=true — enables automatic detection of unproductive agent behavior: consecutive identical tool calls (EXECUTION_MONITOR_SAME_TOOL_LIMIT, default 5) and excessive exploration (EXECUTION_MONITOR_TOTAL_TOOL_LIMIT, default 10) trigger automatic mentor intervention to redirect the agent.

    Reworked Langfuse Observability

    The Langfuse integration has been significantly overhauled for clearer visualization of agent activity. Observation types are now separated into Spans, Generations, Agents, Tools, Chains, Retrievers, Evaluators, Embeddings, and Guardrails. Each agent's tool calls, LLM calls, and intermediate results are tracked as distinct observations, making it straightforward to trace why an agent made a particular decision, what input it received, and what output it produced. Score metrics, timing data, and variable tracking have been improved across all observation types.

    New Capabilities

    Docker Host Network Mode

    Setting DOCKER_NETWORK=host instructs PentAGI to create worker containers using the host network stack instead of a bridge network. This gives containers direct access to local network interfaces, which is necessary for OOB (out-of-band) attack techniques that require binding to local interfaces — such as setting up reverse shells where the listener must be reachable from the target network. Agent prompts include mandatory guidance on OOB port allocation for this mode.

    HTTP Client Timeout

    A new environment variable HTTP_CLIENT_TIMEOUT (default: 600 seconds / 10 minutes) applies a timeout to all outbound HTTP connections — including every LLM provider, search tool, and external API call. Previously, connections to unresponsive backends could hang indefinitely, blocking agent goroutines. When the config is nil, a client with the default timeout is returned instead of Go's http.DefaultClient (which has no timeout).

    Agent Tool Call Limits

    Hard limits are now enforced on the number of tool calls per agent invocation:

    • MAX_GENERAL_AGENT_TOOL_CALLS (default: 100) — applies to primary and specialist agents
    • MAX_LIMITED_AGENT_TOOL_CALLS (default: 20) — applies to focused agents (reflector, mentor, etc.)

    When the limit is reached, the agent is guided to a graceful completion using barrier tools rather than being abruptly terminated.

    Tool Call ID Generation

    A configurable tool call ID template mechanism has been added for LLM backends that require a specific tool call ID format. This prevents validation errors from providers with strict ID format requirements and is set per provider configuration.

    vLLM Reference Configuration

    A tested reference configuration for qwen3.5-27b under vLLM is included, aimed at fully air-gapped or isolated environments where cloud LLM providers are unavailable. The configuration covers model parameters optimized for the agent roles PentAGI uses.

    Flow Templates

    Flows can now be saved as templates and reused. The templates management interface is available in the frontend, with full GraphQL API support for creating, updating, and launching flows from templates.

    Novita AI Provider (Optional)

    Novita AI is available as an optional provider via custom provider YAML configuration (novita.provider.yml). The default model assignment uses moonshotai/kimi-k2.5 for primary agents.

    LLM Provider Improvements

    Updated Model Configurations

    All built-in provider configurations have been updated to reflect the current model landscape:

    • OpenAI: updated to GPT-5.4 series with revised pricing and token limits
    • Anthropic: increased max_tokens limits and updated to latest Claude Sonnet/Opus variants
    • Gemini: updated model assignments including Gemini 2.5-class models
    • Bedrock: added support for Default AWS SDK credential chain, Bearer token, and static credentials (Access Key + Secret Key), in addition to the existing session token method

    Function Call and Thinking Signatures

    Provider-level support for function call signatures and thinking signatures ensures that reasoning-capable models (Claude extended thinking, Gemini thinking tokens, DeepSeek R1 reasoning mode) produce well-formed conversations that preserve reasoning context across multi-turn interactions. This is particularly important for chain summarization, which now retains thinking signatures when compressing long conversations.

    Improved Token Caching

    Token caching has been further optimized for Anthropic (ephemeral cache controls) and Gemini (pre-created content caching) to reduce costs in long-running flows. Cache hit and cache write token counts are tracked separately per turn and are visible in the analytics dashboard.

    Bug Fixes

    Bedrock Provider Compatibility

    Two distinct Bedrock issues have been resolved:

    • Fixed ValidationException errors when using the Converse API with tool schemas generated by Go's jsonschema reflector. The $schema field is now automatically stripped from tool parameters before sending to Bedrock.
    • Fixed a runtime failure where toolConfig was undefined for message chains containing toolUse/toolResult blocks. WithTools is now applied last in both CallEx and CallWithTools to prevent provider config options from overwriting tool definitions.

    Detached Terminal Command Hangs

    Background (detach mode) terminal commands previously inherited the parent agent context. When the parent context was cancelled (e.g., due to agent delegation timeout), the background goroutine was also terminated. Fixed by using context.WithoutCancel for detached goroutines, which preserves context values (tracing, logging) while preventing parent cancellation propagation. The command's own timeout continues to work as expected.

    Infinite Agent Loop Prevention

    Two complementary safeguards are now in place:

    • A hard cap of 100 iterations on the main agent chain loop prevents infinite execution when a model repeatedly calls the same tool.
    • After 3 consecutive identical tool calls, the agent receives a soft "please try another approach" message. After 7 identical consecutive calls in total, the loop terminates with an error.
    • Infinite reflector recursion has been independently fixed by moving retry guards to reflector entry points.

    Logging and Log Worker Fixes

    Several issues with agent activity logging were resolved: fixed log update propagation in the flow assistant log worker, corrected empty log entries being created when no updates occurred during assistant processing, improved streaming log throttling to prevent excessive cache updates, and fixed log worker initialization in assistant mode (missing agent call limits and execution monitoring).

    QA Summarization Double-Summarization

    Fixed an issue where already-summarized sections could be summarized a second time, producing degraded summaries in long-running flows with many completed phases.

    Search Tool HTTP Client Safety

    tavily.go and traversaal.go were mutating Go's global http.DefaultClient.Transport when configuring proxy settings, creating a data race for concurrent requests. Both tools now create a new http.Client instance when a proxy is configured.

    Browser Tool Screenshot Handling

    A screenshot failure in the browser tool no longer discards successfully-fetched page content. The screenshot is treated as a non-critical side effect; on failure, a warning is logged and the page content is returned with an empty screenshot reference.

    Google Search Proxy Configuration

    The Google search tool was constructing a proxy-configured options slice but then ignoring it and using a hardcoded option.WithAPIKey in the actual service creation call. The proxy configuration is now correctly applied.

    User-Defined Provider Precedence

    Fixed an issue (#220) where built-in provider configurations could override user-defined custom configurations with the same provider identifier. User-defined providers now always take precedence.

    Security: CA Private Key Cleanup

    After the server certificate is signed during container startup, the CA private key, CSR, and serial file are now immediately removed from disk. These files are not needed at runtime and their presence increases the attack surface if the container filesystem is compromised.

    Auth Session Management

    Improved session handling in the frontend: WebSocket connections on public pages are prevented, 401/403 errors in WebSocket, GraphQL, and HTTP requests trigger automatic session refresh, and the OAuth providers list is always fetched fresh on the login page to reflect newly added providers without cache clearing.

    Test Coverage

    Backend test coverage has been significantly expanded in this release. Key package coverage after new tests:

    Package Coverage
    pkg/terminal 83.3%
    pkg/queue 89.4%
    pkg/schema 86.5%
    pkg/server/auth 87.7%
    pkg/server/response 100.0%
    pkg/server/context 100.0%
    pkg/csum 84.0%
    pkg/cast 87.3%
    pkg/config 75.7%
    pkg/providers/bedrock 81.1%
    pkg/providers/custom 79.2%
    pkg/providers/embeddings 74.0%
    pkg/providers/deepseek 71.4%
    pkg/providers/glm 71.4%
    pkg/providers/kimi 71.4%
    pkg/providers/qwen 71.4%
    pkg/providers/tester 78.7%

    Tests also cover agent context management, tool registry completeness, executor helpers, terminal formatting utilities, langfuse helpers and noop observer, graphiti disabled mode, server/models validation, and JSON Schema validation.

    Infrastructure

    • Docker Compose healthcheck: pg_isready healthcheck added to the pgvector service so the application waits for the database to be fully ready before starting.
    • License compliance: CONTRIBUTING.md with license compliance guidelines for contributors, and Dockerfile tooling to generate frontend and backend dependency license reports.
    • Frontend terminal: Modular architecture refactor, Unicode rendering fix, and security hardening.

    Documentation

    • Added CONTRIBUTORS.md recognizing all contributors across the full project history (see note below).
    • Updated README with Ollama Cloud setup instructions (Free Tier and Paid Tier).
    • Added reference documentation for Docker host network mode and OOB attack scenarios.
    • Documented agent supervision settings (AGENT_PLANNING_STEP_ENABLED, EXECUTION_MONITOR_ENABLED, MAX_GENERAL_AGENT_TOOL_CALLS, MAX_LIMITED_AGENT_TOOL_CALLS, HTTP_CLIENT_TIMEOUT).

    A Note on Repository History

    On March 29, 2026, the repository history prior to that date was rewritten into a single squash commit to resolve a licensing matter. Individual commit history from January 2025 through March 2026 is no longer visible in the GitHub interface. The CONTRIBUTORS.md file was created to permanently record all contributions made during that period.

    Contributors

    Core Team

    • @asdek (Dmitry Nagibin) — Architecture, backend infrastructure, agent system, provider integrations, observability, project coordination
    • @sirozha (Sergey Kozyrenko) — React UI, settings interfaces, GraphQL integration, frontend architecture, analytics dashboard, terminal component
    • @zavgorodnii (Andrei Zavgorodnii) — Graphiti integration, patch refiner, knowledge graph implementation

    External Contributors

    Full Changelog: https://github.com/vxcontrol/pentagi/compare/v1.2.0...v2.0.0

    Downloads
  • v1.2.0 b90ea4711e
    Compare

    Stable

    yindo released this 2026-03-25 23:16:07 -04:00 | 356 commits to main since this release

    📅 Originally published on GitHub: Wed, 25 Feb 2026 16:26:25 GMT
    🏷️ Git tag created: Thu, 26 Mar 2026 03:16:07 GMT

    🚀 PentAGI 1.2 - Enhanced AI Capabilities! Major upgrade bringing latest reasoning models, token caching, comprehensive analytics, and REST API access for seamless integration with automation platforms.

    Discord⠀Telegram

    🎯 Major Features

    🧠 Latest Reasoning Models Support - Complete integration of cutting-edge AI models with native reasoning capabilities:

    • Gemini 2.5/3.0 family with thinking tokens support
    • Anthropic Claude Sonnet 4+ with extended reasoning
    • DeepSeek R1 and Kimi K2.5 in reasoning mode
    • OpenAI o-series models with signature thoughts
    • OpenRouter and OpenAI-compatible endpoints with reasoning content preservation

    💰 Token Caching & Cost Optimization - Intelligent prompt caching reduces input token costs by 40-70% in multi-turn agent conversations:

    • Native caching support for Anthropic (ephemeral cache controls) and Gemini (pre-created content caching)
    • Automatic cache hit tracking with detailed analytics
    • Particularly effective for long-context penetration testing sessions
    • Standardized cache token reporting across all providers

    📊 Usage Analytics & Monitoring - Comprehensive REST API endpoints for detailed resource utilization tracking:

    • Token usage breakdown by agent type (researcher/developer/executor)
    • Cost analysis with cache read/write separation
    • Execution time metrics per flow and subtask
    • Tool call frequency statistics
    • Foundation for visual analytics dashboard (coming in v1.3)

    🔑 API Token Management - JWT-based API authentication enables programmatic access to PentAGI:

    • Generate and manage API tokens through web interface
    • Full REST and GraphQL API access for automation
    • OpenAPI specifications for client code generation in any language
    • Integration-ready for n8n, OpenClaw, Claude Desktop, and custom solutions
    • Foundation for official MCP server (planned for future releases)

    🔍 Sploitus Integration - Experimental support for vulnerability search engine:

    • Cloudflare-protected service requires IP reputation verification
    • Use built-in ftester utility to check your IP reputation before enabling
    • Configure via SPLOITUS_ENABLED environment variable

    📡 Langfuse v3 Observability - Complete migration to Langfuse v3 standard with enhanced LLM operations tracking:

    • Observation type separation: Spans, Generations, Agents, Tools, Chains, Retrievers, Evaluators, Embeddings, Guardrails
    • Enhanced message chain visualization with Playground mode navigation
    • Detailed Score metrics and execution time logging
    • Improved variable and metadata tracking across all observation types

    🚀 New Features

    • Reasoning Content Preservation: Smart message chain summarization that maintains reasoning signatures for models requiring strict conversation structure
    • Tool Call ID Templates: Configurable tool call ID format enforcement for LLM backends with strict validation requirements
    • User Preferences System: Favorite flows management with persistent preferences storage
    • GraphQL Subscriptions: Real-time flow updates with user-specific event publishing
    • Docker Build Versioning: Embedded version and revision information in container images with dedicated build scripts for Linux/macOS/Windows
    • Enhanced Error Diagnostics: Stop reason included in error messages (e.g., length indicates need to increase max_tokens)
    • PDF Report Generation: Export flow results to PDF using @react-pdf/renderer library
    • User Favorites: Add and manage favorite flows with dedicated GraphQL mutations
    • Podman Support: Official documentation for running PentAGI with Podman in rootless mode

    🎨 UI/UX Improvements

    • Enhanced Theme Handling: Improved dark/light/system theme switching with automatic system preference detection
    • Better Authentication Flow: Safe return URL handling with validation to prevent open redirect vulnerabilities
    • Google OAuth Fix: Resolved CORS issues and improved cookie handling for Google OAuth integration
    • Flow Subscriptions: Real-time flow updates in UI via GraphQL subscriptions with user-scoped events
    • Settings Form Validation: Stronger password requirements with visibility toggles
    • Enhanced Report Generation: Fixed markdown rendering issues in flow reports

    🐛 Key Fixes

    • Resource Leak Prevention: Fixed response body leaks in browser tool, added tar header size validation in terminal operations, properly close tarWriter to prevent incomplete archives (#101)
    • Security Hardening:
      • OAuth state parameter validation with explicit CSRF checks (#101)
      • Session expiry enforcement in authentication middleware
      • SameSite cookie attributes for CSRF protection
      • Browser tool HTTP client timeout (30s) to prevent indefinite hangs
      • Authorization string typos fixed (trailing quotes causing ACL failures)
    • TLS Configuration: Respect EXTERNAL_SSL_INSECURE config in Langfuse client, load custom CA certificates from EXTERNAL_SSL_CA_PATH, use system cert pool as base (#132)
    • Terminal Command Logic: Corrected terminal command handling logic (#124)
    • Swagger Documentation: Fixed missing closing quotes in OpenAPI annotations
    • Code Quality: Removed debug console.log statements from production code
    • Traversaal API: Updated integration after vendor-side API specification changes
    • Nil Pointer Checks: Added nil checks for Langfuse client before ForceFlush operations

    🔧 Infrastructure Improvements

    • LangChainGo v0.1.14-update.1: Major dependency update with 6 months of accumulated improvements:
      • Signature thoughts support for Anthropic, Gemini, OpenAI providers
      • Message chain caching for Gemini and Anthropic with token savings tracking
      • Standardized usage format across all providers with unified field names
      • Comprehensive test coverage for LLM scenarios including multi-turn conversations, function calling, caching validation
      • Migrated Google AI provider to google.golang.org/genai from deprecated SDK
      • Bedrock Converse API support for Anthropic Claude models
      • Enhanced streaming with proper resource cleanup (memory leak fixes)
    • Alpine 3.23.3: Updated base Docker image with latest security patches
    • Model Updates: Switched from deprecated gemini-2.0-flash-lite to gemini-2.5-flash-lite with adjusted pricing
    • GitHub Actions Modernization: Upgraded all workflows for Node 24 compatibility
    • Dependency Security Updates:
      • axios 1.13.2 → 1.13.5
      • lodash 4.17.21 → 4.17.23
      • diff 5.2.0 → 5.2.2
      • jspdf 4.1.0 → 4.2.0
    • External Network Access: Comprehensive documentation for configuring PentAGI accessibility from other machines with firewall setup instructions
    • Entrypoint Script: SSL certificate generation management for enhanced security setup

    🔄 Performance & Architecture

    • Standardized Token Usage: All LLM providers now return consistent token fields (PromptTokens, CompletionTokens, TotalTokens, CacheCreationTokens, CacheReadTokens)
    • Enhanced Logging: Enriched log fields with flow/task/subtask IDs for better traceability
    • Observation Framework: Refactored observability with W3C Trace Context compliance (newSpanID/newTraceID functions)
    • Chain Summarization: Enhanced algorithm with critical guarantees preserving last N QA sections even if exceeding size limits, ensuring reasoning signatures retention
    • Improved Metadata Handling: Stop reason tracking in generation metadata for better observability

    📚 Documentation

    • Typo Fixes: Comprehensive typo corrections across documentation and code comments (#121):
      • "PegtAGI" → "PentAGI" in frontend README
      • "Depp Infra" → "Deep Infra", "Traversal" → "Traversaal" in EULA
      • OAuth environment variable names aligned with .env.example
      • Fixed filename typos (sreenshots.go → screenshots.go, wizard-integation → wizard-integration)
    • External Access Guide: Step-by-step instructions for PENTAGI_LISTEN_IP, PUBLIC_URL, CORS_ORIGINS configuration
    • Podman Documentation: Running PentAGI with Podman in rootless mode with non-privileged ports

    📖 Documentation: For detailed setup instructions, visit the README and Quick Start Guide

    New Contributors

    Full Changelog: https://github.com/vxcontrol/pentagi/compare/v1.1.0...v1.2.0

    Downloads
  • v1.1.0 b90ea4711e
    Compare

    Stable

    yindo released this 2026-03-25 23:16:07 -04:00 | 356 commits to main since this release

    📅 Originally published on GitHub: Sat, 17 Jan 2026 23:30:43 GMT
    🏷️ Git tag created: Thu, 26 Mar 2026 03:16:07 GMT

    🔧 Bug Fixes & Improvements

    LiteLLM Passthrough Support

    • Fixed Gemini provider compatibility issues preventing proper LiteLLM integration
    • All providers now support LiteLLM passthrough mode with standardized endpoints:
      • OpenAI: http://litellm:4000/openai/v1
      • Anthropic: http://litellm:4000/anthropic/v1
      • Gemini: http://litellm:4000/gemini
    • Tested and verified with LiteLLM v1.80.11-stable.1
    • Enhanced Gemini provider with custom HTTP transport for API key injection and URL rewriting

    Windows File Path Compatibility

    • Changed file mounting scheme in PentAGI container to resolve Windows path format issues
    • Migrated from host path mapping to fixed container paths for better cross-platform compatibility
    • Updated volume mounts:
      • PENTAGI_LLM_SERVER_CONFIG_PATH → /opt/pentagi/conf/custom.provider.yml
      • PENTAGI_OLLAMA_SERVER_CONFIG_PATH → /opt/pentagi/conf/ollama.provider.yml
      • PENTAGI_DOCKER_CERT_PATH → /opt/pentagi/docker/ssl
    • Migration: Installer v1.0.0 automatically migrates old settings to new schema
    • Users can now specify absolute paths in their host filesystem through installer forms

    Ollama Single Model Configuration

    • Added OLLAMA_SERVER_MODEL environment variable to select a single model for all agents
    • Eliminates need to create custom provider configuration files for simple setups
    • Additional fine-tuning options:
      • OLLAMA_SERVER_PULL_MODELS_ENABLED - Control automatic model downloads (default: false)
      • OLLAMA_SERVER_LOAD_MODELS_ENABLED - Query available models on startup (default: false)
      • OLLAMA_SERVER_PULL_MODELS_TIMEOUT - Timeout for model pulls in seconds (default: 600)
    • Default model: llama3.1:8b-instruct-q8_0

    Installer v1.0.0

    • Bumped installer version to 1.0.0 with comprehensive stability improvements
    • Full Windows support with all configuration scenarios matching Linux and macOS
    • Disabled ANSI formatting in Docker Compose commands on Windows for cleaner console output
    • Automatic settings migration from old path variables to new schema:
      • DOCKER_CERT_PATH → PENTAGI_DOCKER_CERT_PATH
      • LLM_SERVER_CONFIG_PATH → PENTAGI_LLM_SERVER_CONFIG_PATH
      • OLLAMA_SERVER_CONFIG_PATH → PENTAGI_OLLAMA_SERVER_CONFIG_PATH
    • Enhanced error handling and validation throughout installation process
    • Recommended: Download latest installer, run "Apply changes" to migrate to new file mounting scheme, then navigate to Maintenance tab and execute "Update PentAGI" to download the new version that supports these options

    Enhanced Terminal Command Handling

    • Improved agents' understanding of empty results from synchronous terminal commands
    • Enhanced background command processing with asynchronous result capture
    • Introduced quick check timeout for background command execution
    • Clearer feedback on command failures and silent successes
    • More accurate success messages reflecting actual command execution outcomes

    Additional Improvements

    • Fixed nil pointer dereference in Graphiti client methods
    • Enhanced error handling for invalid server and proxy URLs during provider initialization
    • Improved Docker Compose command handling on Windows systems
    • Added comprehensive tests for Gemini API key injection and URL rewriting

    Full Changelog: https://github.com/vxcontrol/pentagi/compare/v1.0.1...v1.1.0

    Downloads
  • v1.0.1 b90ea4711e
    Compare

    Stable

    yindo released this 2026-03-25 23:16:07 -04:00 | 356 commits to main since this release

    📅 Originally published on GitHub: Tue, 06 Jan 2026 17:53:27 GMT
    🏷️ Git tag created: Thu, 26 Mar 2026 03:16:07 GMT

    🐛 Bug Fixes & Improvements

    Enhanced Error Diagnostics

    • Added stop reason to error messages when LLM fails to generate tool calls
    • If stop reason is length, increase max_tokens parameter for the affected agent in provider settings
    • Improves troubleshooting and configuration optimization

    DuckDuckGo Search Stability

    • Migrated to new DuckDuckGo API with HTML response parsing
    • Added comprehensive test coverage with real-world search scenarios
    • Significantly improved reliability and result quality

    Provider Guardrails Bypass

    • Added explicit authorization framework to all agent prompts
    • Prevents blocking by OpenAI, Anthropic, and Google Gemini content filters
    • Clarified penetration testing context as pre-authorized activity

    OpenAI Configuration Updates

    • Temporarily switched from gpt-5 to o4-mini for primary agent and assistant due to OpenAI prompt evaluation instability
    • Increased max_tokens limits across multiple agents for better output capacity
    • Recommendation: Enable Human-in-the-loop mode (ASK_USER=true in .env) when using OpenAI provider for improved stability

    Additional Improvements

    • Enhanced message formatting in vector store communications with document match scores
    • Improved clarity in generator and refiner prompts for user task interpretation
    • Added customer interaction protocol for AskUser tool

    Full Changelog: https://github.com/vxcontrol/pentagi/compare/v1.0.0...v1.0.1

    Downloads
  • v1.0.0 b90ea4711e
    Compare

    Stable

    yindo released this 2026-03-25 23:16:07 -04:00 | 356 commits to main since this release

    📅 Originally published on GitHub: Wed, 31 Dec 2025 11:15:01 GMT
    🏷️ Git tag created: Thu, 26 Mar 2026 03:16:07 GMT

    🎉 PentAGI 1.0 - Production Ready! The first stable release of our autonomous penetration testing platform, bringing enterprise-grade features, enhanced AI capabilities, and a completely redesigned user experience.

    Discord⠀Telegram

    🎯 Major Features

    🧠 Graphiti Knowledge Graph Integration - Revolutionary memory system using Graphiti, a temporal knowledge graph that maintains context across penetration testing sessions. Configure deployment modes (embedded, external, or disabled) and leverage graph-based reasoning for more intelligent agent decisions.

    ⚙️ Interactive Installer - Professional setup wizard with comprehensive system checks, Docker volume detection, and step-by-step configuration for all services including LLM providers, search engines, and observability stack. Available for Linux, macOS, and Windows.

    🎨 Modern Frontend Redesign - Complete UI/UX overhaul with React 19, Tailwind CSS v4, and enhanced architecture:

    • Advanced flow management with filters by agents, tasks, tools, and vector stores
    • Improved settings interface with tabular data views and form validation
    • Real-time toast notifications and responsive design
    • Enhanced favorites system and sidebar navigation

    🔧 Provider Management System - Unified configuration for LLM providers with support for:

    • AWS Bedrock (with temporary credentials and session tokens)
    • Google Gemini (2.5 Flash and Pro models)
    • Ollama (local deployment)
    • Custom OpenAI-compatible endpoints

    🔍 Enhanced Search Ecosystem - Integrated SearXNG meta-search engine with privacy-focused searching, complementing existing Perplexity and DuckDuckGo providers.

    ⚡ Patch Refiner - Intelligent result refinement system that automatically improves agent outputs, validates findings, and ensures accuracy before presenting final results.

    🚀 New Features

    • Prompt & Agent Management: Create, edit, and test custom AI agent configurations through the web interface
    • Provider Testing UI: Built-in testing functionality for validating LLM provider configurations with detailed reports
    • SSL/TLS Configuration: External certificate support with custom CA paths and insecure mode for development
    • Enhanced Container Management: Configurable Docker images for penetration testing with improved isolation
    • Installation ID & Licensing: PentAGI Cloud API integration with license key management
    • Volume Persistence Detection: Automatic Docker volume existence checks for Pentagi and Langfuse services

    🎨 UI/UX Improvements

    • React 19 Migration: Upgraded to latest React version with improved performance and new features
    • Tailwind v4: Modern styling system with better customization and smaller bundle size
    • Flow Components Refactoring: Renamed and restructured chat components to flow-based architecture
    • Responsive Filter Forms: Advanced filtering UI with input groups for agents, tools, and vector stores
    • Delete Confirmation Dialogs: User-friendly confirmation prompts for destructive actions
    • Hybrid Model Selector: Combined input/dropdown for flexible model selection

    🐛 Key Fixes

    • Long Subtask Descriptions: Fixed issue #72 with truncating overly detailed subtask descriptions
    • Large Result Storage: Resolved database issues when storing extensive task and subtask results
    • Provider Configuration: Fixed empty provider config creation in web UI (issue #63)
    • Manual Task Stopping: Improved handling when stopping tasks during refiner operations
    • URL Resolution: Enhanced browser tool logic for local domain handling
    • UTF-8 Sanitization: Moved database UTF-8 sanitization to common package for consistency

    🔧 Infrastructure Improvements

    • Go 1.24: Updated to latest Golang version with performance improvements
    • Docker Optimization: Removed unnecessary cleanup steps, optimized layer caching
    • Dependency Updates:
      • Docker SDK 28.2.2 → 28.3.3
      • Ollama 0.9.6 → 0.10.0
      • golang.org/x/crypto security updates
      • Frontend dependencies (Vite, axios, jsPDF)
    • Alpine 3.22.1: Latest base image with security patches
    • Enhanced Logging: Improved logging throughout the application with better context

    🔄 Performance & Architecture

    • Graphite Docker Compose: Separate docker-compose-graphiti.yml for modular deployment
    • Parallel Workers: Increased default from 8 to 16 for improved testing performance
    • Apollo Cache Optimization: Added keyFields normalization for better provider handling
    • Agent Configuration: Refactored to use AgentConfigType enum for clarity
    • Settings Refactoring: Streamlined provider, prompt, and agent management architecture

    📚 Documentation

    • Installer Documentation: Comprehensive guide for installation wizard components
    • Flow Execution Guide: Enhanced launch configuration documentation
    • Graphiti Setup: Instructions for running knowledge graph stack
    • Provider Examples: Updated configuration examples for all supported providers
    • Prerequisites Guide: Detailed Docker installation and permissions documentation

    📖 Documentation: For detailed setup instructions, visit the README and Quick Start Guide

    New Contributors

    Full Changelog: https://github.com/vxcontrol/pentagi/compare/v0.3.0...v1.0.0

    Downloads
  • v0.3.0 b90ea4711e
    Compare

    Stable

    yindo released this 2026-03-25 23:16:07 -04:00 | 356 commits to main since this release

    📅 Originally published on GitHub: Wed, 25 Jun 2025 22:07:13 GMT
    🏷️ Git tag created: Thu, 26 Mar 2026 03:16:07 GMT

    🚀 Join the Community! Connect with security researchers, AI enthusiasts, and fellow ethical hackers. Get support, share insights, and stay updated with the latest PentAGI developments.

    Discord⠀Telegram

    🎯 Major Features

    🤖 Assistant Mode - Complete interactive AI assistant with streaming responses, persistent chat sessions, and intelligent agent delegation. Create multiple chat sessions and seamlessly switch between manual assistance and automated penetration testing workflows.

    🧪 Professional Testing Suite - Three specialized testing utilities:

    • ctester: Test LLM agent configurations with parallel execution and detailed reporting
    • etester: Manage vector embeddings with provider testing and database optimization
    • ftester: Debug individual functions and AI behaviors with interactive mock modes

    🔍 Enhanced Search Capabilities - Integrated Perplexity AI and DuckDuckGo search engines alongside existing providers, plus multi-provider embedding system supporting OpenAI, Ollama, Mistral, Jina, HuggingFace, GoogleAI, and VoyageAI.

    🛡️ Custom Kali Linux Environment - Dedicated Docker image optimized for penetration testing with enhanced security tools and network admin capabilities. The open-source build configuration is available under MIT license with automated multi-platform builds and security attestations.

    ⚡ Enhanced LLM Integration - PentAGI now uses a custom fork of langchaingo with significant improvements for better LLM provider compatibility, enhanced function calling, streaming responses, and optimized external service integrations.

    🚀 New Features

    • Community Launch: Official Discord and Telegram channels for community support, knowledge sharing, and collaboration between security researchers and AI enthusiasts
    • Flexible LLM Configuration: YAML/JSON configuration system for custom providers with per-agent model specifications (examples)
    • Advanced Report Generation: Comprehensive Markdown and PDF reports for flows, tasks, and subtasks
    • Smart Context Management: Enhanced conversation summarization with configurable preservation settings
    • Message Copy & Search: Copy messages in Markdown format with text highlighting across all interfaces
    • Provider Management: Visual icons and improved status indicators for OpenAI, Anthropic, and custom providers

    🎨 UI/UX Improvements

    • Streamlined Assistant Interface: New tab with chat creation, management, and persistent state
    • Enhanced Navigation: Improved breadcrumbs with status and provider information
    • Better Authentication: Enhanced GitHub/Google OAuth with password change functionality
    • Improved Flow Management: Better status handling with proper state transitions and input blocking
    • Professional Tooltips: Fixed positioning and enhanced visual feedback

    🐛 Key Fixes

    • Flow Status Synchronization: Resolved issues with status updates when switching between flows
    • Assistant Integration: Fixed problems launching assistants on new and completed flows
    • Terminal Synchronization: Improved command execution display between automated and manual agents
    • Message Chain Consistency: Enhanced restoration and context handling after interruptions
    • Configuration Issues: Resolved Docker, environment variables, and provider setup problems

    🔧 Infrastructure Improvements

    • Enhanced Container Security: Improved isolation with controlled network capabilities
    • Environment Flexibility: ASK_USER interactive mode, proxy support, and SSL/TLS enhancements
    • Build Optimization: Golang 1.24 upgrade, dependency updates, and improved Docker builds
    • Configuration Management: Pre-built provider configs for OpenRouter, DeepInfra, and DeepSeek
    • Custom Docker Images: Open-source Kali Linux containers with automated builds, multi-platform support, and security attestations

    🔄 Performance & Architecture

    • Agent System Refactoring: Major improvements to core execution logic with better modularity
    • Memory Optimization: Enhanced context management and chain summarization for reduced footprint
    • Database Performance: Optimized queries and improved vector storage operations
    • Enhanced Prompt System: Unified templates with shared components and simplified handling
    • LLM Library Improvements: Migration to custom langchaingo fork with enhanced streaming, function calling, and provider compatibility

    📖 Documentation: For detailed setup instructions, visit the README and Quick Start Guide

    New Contributors

    Full Changelog: https://github.com/vxcontrol/pentagi/compare/v0.2.0...v0.3.0

    Downloads
  • v0.2.0 b90ea4711e
    Compare

    Pre-Release

    yindo released this 2026-03-25 23:16:07 -04:00 | 356 commits to main since this release

    📅 Originally published on GitHub: Thu, 09 Jan 2025 13:55:06 GMT
    🏷️ Git tag created: Thu, 26 Mar 2026 03:16:07 GMT

    🚀 New Features

    Frontend Architecture

    • ✨ Implemented modern React 18 architecture with TypeScript for enhanced type safety
    • 🎨 Added shadcn/ui components with Radix UI primitives for consistent design
    • 🌓 Introduced dark/light theme support with Tailwind CSS
    • 📱 Added responsive design support for mobile, tablet, and desktop layouts
    • ⚡ Optimized build process with Vite and module chunking

    Core Features

    • 💬 Real-time chat interface with AI agents using WebSocket subscriptions
    • 🤖 Multi-agent system with specialized roles (Researcher, Developer, Executor)
    • 📊 Terminal integration with real-time output monitoring
    • 🎯 Task tracking system with subtasks and progress monitoring
    • 🔍 Integrated search capabilities with vector store
    • 📸 Screenshot capture and management system

    Security & Authentication

    • 🔐 Multi-provider authentication support
    • 🔑 OAuth integration with GitHub and Google
    • 🛡️ SSL/TLS support for secure communications
    • 🔒 Environment-based configuration management

    🐛 Bug Fixes

    • Fixed WebSocket connection handling for GraphQL subscriptions
    • Improved error handling in terminal output
    • Resolved theme switching persistence issues
    • Fixed mobile layout responsiveness

    🔄 Changes

    • Migrated from CRA to Vite for better build performance
    • Updated all dependencies to latest stable versions
    • Improved code organization with feature-based structure
    • Enhanced type definitions for better TypeScript support

    📚 Documentation

    • Added comprehensive frontend documentation
    • Included development setup instructions
    • Added component architecture documentation
    • Updated environment configuration guide

    🛠️ Technical Details

    • React 18.3.1
    • TypeScript 5.6.2
    • Vite 5.4.7
    • GraphQL 16.9.0
    • Tailwind CSS 3.4.13

    🔜 Coming Soon

    • Enhanced performance monitoring
    • Improved error reporting
    • Extended test coverage
    • Additional UI components

    🙏 Acknowledgments

    • Thanks to @sirozha for new frontend version

    What's Changed

    New Contributors

    Full Changelog: https://github.com/vxcontrol/pentagi/compare/v0.1.0...v0.2.0

    Downloads
  • v0.1.0 b90ea4711e
    Compare

    Pre-Release

    yindo released this 2026-03-25 23:16:07 -04:00 | 356 commits to main since this release

    📅 Originally published on GitHub: Tue, 07 Jan 2025 01:26:34 GMT
    🏷️ Git tag created: Thu, 26 Mar 2026 03:16:07 GMT

    🎯 Current State

    PentAGI is in early alpha stage, focusing on core functionality and system stability. This release demonstrates the basic capabilities of our autonomous penetration testing system while actively being developed and improved.

    ✨ Available Features

    Core Functionality

    • 🤖 Multi-agent system (Researcher, Developer, Executor)
    • 🛡️ Integration with essential security testing tools
    • 🧠 Basic memory system with vector storage
    • 🔄 Autonomous decision-making capabilities

    Technical Implementation

    • 🐳 Docker-based deployment
    • 📊 Basic monitoring (Grafana + OpenTelemetry)
    • 📝 LLM operations tracking (Langfuse)
    • 🔌 Support for OpenAI/Anthropic APIs

    ⚠️ Important Notes

    • This is an alpha release intended for testing and feedback
    • Not recommended for production use
    • Expect frequent updates and changes
    • Some features may be unstable or incomplete
    • Limited documentation available

    🚀 Quick Start

    mkdir pentagi && cd pentagi
curl -O https://raw.githubusercontent.com/vxcontrol/pentagi/main/docker-compose.yml
curl -o .env https://raw.githubusercontent.com/vxcontrol/pentagi/main/.env.example
# Configure your .env file
docker compose up -d

    For detailed documentation and latest updates, please visit README file.

    Downloads