Files
qwen-code/packages/cua-driver/contract/manifest.json
T
顾盼 a791e47a54 feat(cua): streamline computer use and port native reliability fixes (#14846)
* feat(cua): preload the JavaScript computer use runtime

* refactor(cua): remove the native Wayland backend

* feat(cua): move app computer use into the driver

* feat(cua): preserve background delivery and port platform fixes

* feat(cua): add explicit desktop input and browser runtime facade

* feat(cua): package the unified 0.26.0 computer use runtime

* fix(cua): finalize cross-platform native reliability

* fix(cua): isolate runtime releases from CLI versioning

* fix(cua): keep Electron out of root workspace installs

* fix(cua): synchronize consumers and async cleanup tests

* fix(browser): migrate smoke to preloaded cua runtime

* docs(cua): document MCP setup for Qwen Code Claude Code and Codex
2026-10-21 14:03:31 +08:00

4306 lines
112 KiB
JSON

{
"generated_notice": "Generated by cua-contract-gen; do not edit by hand.",
"experimental": true,
"contract_version": "0.7.0",
"tools_list_schema_version": "1",
"capability_version": "1",
"mcp_protocol_version": "2025-06-18",
"transport": "mcp_stdio",
"tools": [
{
"name": "click",
"description": "Click an absolute point in get_desktop_state coordinates without targeting a window.",
"platforms": [
"macos",
"windows",
"linux"
],
"capabilities": [
"input.pointer.click",
"input.pointer.click.left",
"accessibility.element_tokens"
],
"annotations": {
"read_only": false,
"destructive": true,
"idempotent": false,
"open_world": true
},
"schema_mode": "portable_subset",
"cursor_semantics": {
"action": "click"
},
"input_schema": {
"additionalProperties": false,
"properties": {
"button": {
"enum": [
"left",
"right",
"middle"
],
"type": "string"
},
"count": {
"maximum": 3,
"minimum": 1,
"type": "integer"
},
"scope": {
"description": "Deprecated flat desktop target retained for wire compatibility.",
"enum": [
"desktop"
],
"type": "string"
},
"session": {
"description": "For multi-call work, prefer a short public session label and repeat it on every call that\naccepts it. Omit it to use the authenticated transport's implicit lifecycle session.",
"type": "string"
},
"target": {
"additionalProperties": true,
"description": "Exact window or desktop target. Window requires pid and window_id; desktop requires display_id.",
"properties": {
"display_id": {
"description": "Use primary for the primary display.",
"type": "string"
},
"kind": {
"enum": [
"window",
"desktop"
],
"type": "string"
},
"pid": {
"minimum": 0,
"type": "integer"
},
"window_id": {
"minimum": 0,
"type": "integer"
}
},
"required": [
"kind"
],
"type": "object"
},
"x": {
"type": "number"
},
"y": {
"type": "number"
}
},
"required": [
"x",
"y"
],
"type": "object"
},
"success_output_schema": {
"additionalProperties": false,
"properties": {
"delivery": {
"additionalProperties": false,
"properties": {
"delivered_count": {
"minimum": 0,
"type": [
"integer",
"null"
]
},
"mode": {
"enum": [
"background",
"foreground",
"not_applicable",
"unknown"
],
"type": "string"
}
},
"required": [
"mode"
],
"type": [
"object",
"null"
]
},
"effect": {
"enum": [
"confirmed",
"partial",
"unverifiable",
"suspected_noop",
"refused"
],
"type": "string"
},
"escalation": {
"additionalProperties": false,
"properties": {
"reason": {
"enum": [
"route_unavailable",
"delivery_failed",
"effect_unconfirmed",
"suspected_noop",
"permission_required"
],
"type": "string"
},
"target": {
"enum": [
"pixel",
"foreground",
"page",
"session"
],
"type": "string"
}
},
"required": [
"target",
"reason"
],
"type": [
"object",
"null"
]
},
"evidence": {
"items": {
"additionalProperties": false,
"properties": {
"kind": {
"enum": [
"value_readback",
"window_change"
],
"type": "string"
}
},
"required": [
"kind"
],
"type": "object"
},
"type": [
"array",
"null"
]
},
"route": {
"enum": [
"accessibility",
"synthetic_events",
"global_input",
"system_api",
"dom",
"trusted_input"
],
"type": "string"
}
},
"required": [
"effect",
"route"
],
"type": "object"
}
},
{
"name": "clipboard_read",
"description": "List available system clipboard types and optionally return privacy-sensitive plain text. Clipboard content is never retained in telemetry.",
"platforms": [
"macos",
"windows",
"linux"
],
"capabilities": [
"clipboard.read",
"clipboard.types"
],
"annotations": {
"read_only": true,
"destructive": false,
"idempotent": false,
"open_world": false
},
"schema_mode": "canonical_runtime",
"cursor_semantics": {
"action": "observe"
},
"input_schema": {
"additionalProperties": false,
"properties": {
"include_text": {
"default": false,
"description": "Return plain-text clipboard content in addition to the available types.\nClipboard content is privacy-sensitive and is never retained in telemetry.",
"type": "boolean"
},
"session": {
"description": "For multi-call work, prefer a short public session label and repeat it on every call that\naccepts it. Omit it to use the authenticated transport's implicit lifecycle session.",
"type": "string"
}
},
"required": [],
"type": "object"
},
"success_output_schema": {
"additionalProperties": true,
"properties": {
"content_redacted_from_telemetry": {
"type": "boolean"
},
"privacy_sensitive": {
"type": "boolean"
},
"supported": {
"type": "boolean"
},
"text": {
"anyOf": [
{
"type": "string"
},
{
"type": "null"
}
]
},
"types": {
"items": {
"type": "string"
},
"type": "array"
}
},
"required": [
"supported",
"types",
"text",
"privacy_sensitive",
"content_redacted_from_telemetry"
],
"type": "object"
}
},
{
"name": "clipboard_write",
"description": "Replace the system clipboard with exactly one value: plain text, an image from an absolute local path, or a file URL from an absolute local path. Returns the available types for read-back before paste.",
"platforms": [
"macos",
"windows",
"linux"
],
"capabilities": [
"clipboard.write",
"clipboard.write.text",
"clipboard.write.image",
"clipboard.write.file_url",
"clipboard.types"
],
"annotations": {
"read_only": false,
"destructive": true,
"idempotent": true,
"open_world": false
},
"schema_mode": "canonical_runtime",
"cursor_semantics": {
"action": "text"
},
"input_schema": {
"additionalProperties": false,
"properties": {
"file_path": {
"description": "Absolute path to a local file to place on the clipboard as a file URL.",
"type": "string"
},
"image_path": {
"description": "Absolute path to a local image to place on the clipboard.",
"type": "string"
},
"session": {
"description": "For multi-call work, prefer a short public session label and repeat it on every call that\naccepts it. Omit it to use the authenticated transport's implicit lifecycle session.",
"type": "string"
},
"text": {
"description": "Plain text to place on the clipboard.",
"type": "string"
}
},
"required": [],
"type": "object"
},
"success_output_schema": {
"additionalProperties": true,
"properties": {
"content_redacted_from_telemetry": {
"type": "boolean"
},
"privacy_sensitive": {
"type": "boolean"
},
"supported": {
"type": "boolean"
},
"types": {
"items": {
"type": "string"
},
"type": "array"
},
"written_type": {
"type": "string"
}
},
"required": [
"supported",
"written_type",
"types",
"privacy_sensitive",
"content_redacted_from_telemetry"
],
"type": "object"
}
},
{
"name": "drag",
"description": "Drag between two absolute points in get_desktop_state coordinates.",
"platforms": [
"macos",
"windows",
"linux"
],
"capabilities": [
"input.pointer.drag"
],
"annotations": {
"read_only": false,
"destructive": true,
"idempotent": false,
"open_world": true
},
"schema_mode": "portable_subset",
"cursor_semantics": {
"action": "drag"
},
"input_schema": {
"additionalProperties": false,
"properties": {
"button": {
"enum": [
"left",
"right",
"middle"
],
"type": "string"
},
"duration_ms": {
"maximum": 10000,
"minimum": 0,
"type": "integer"
},
"from_x": {
"type": "number"
},
"from_y": {
"type": "number"
},
"modifier": {
"items": {
"type": "string"
},
"type": "array"
},
"scope": {
"description": "Deprecated flat desktop target retained for wire compatibility.",
"enum": [
"desktop"
],
"type": "string"
},
"session": {
"description": "For multi-call work, prefer a short public session label and repeat it on every call that\naccepts it. Omit it to use the authenticated transport's implicit lifecycle session.",
"type": "string"
},
"steps": {
"maximum": 200,
"minimum": 1,
"type": "integer"
},
"target": {
"additionalProperties": true,
"description": "Exact window or desktop target. Window requires pid and window_id; desktop requires display_id.",
"properties": {
"display_id": {
"description": "Use primary for the primary display.",
"type": "string"
},
"kind": {
"enum": [
"window",
"desktop"
],
"type": "string"
},
"pid": {
"minimum": 0,
"type": "integer"
},
"window_id": {
"minimum": 0,
"type": "integer"
}
},
"required": [
"kind"
],
"type": "object"
},
"to_x": {
"type": "number"
},
"to_y": {
"type": "number"
}
},
"required": [
"from_x",
"from_y",
"to_x",
"to_y"
],
"type": "object"
},
"success_output_schema": {
"additionalProperties": false,
"properties": {
"delivery": {
"additionalProperties": false,
"properties": {
"delivered_count": {
"minimum": 0,
"type": [
"integer",
"null"
]
},
"mode": {
"enum": [
"background",
"foreground",
"not_applicable",
"unknown"
],
"type": "string"
}
},
"required": [
"mode"
],
"type": [
"object",
"null"
]
},
"effect": {
"enum": [
"confirmed",
"partial",
"unverifiable",
"suspected_noop",
"refused"
],
"type": "string"
},
"escalation": {
"additionalProperties": false,
"properties": {
"reason": {
"enum": [
"route_unavailable",
"delivery_failed",
"effect_unconfirmed",
"suspected_noop",
"permission_required"
],
"type": "string"
},
"target": {
"enum": [
"pixel",
"foreground",
"page",
"session"
],
"type": "string"
}
},
"required": [
"target",
"reason"
],
"type": [
"object",
"null"
]
},
"evidence": {
"items": {
"additionalProperties": false,
"properties": {
"kind": {
"enum": [
"value_readback",
"window_change"
],
"type": "string"
}
},
"required": [
"kind"
],
"type": "object"
},
"type": [
"array",
"null"
]
},
"route": {
"enum": [
"accessibility",
"synthetic_events",
"global_input",
"system_api",
"dom",
"trusted_input"
],
"type": "string"
}
},
"required": [
"effect",
"route"
],
"type": "object"
}
},
{
"name": "end_session",
"description": "End one visible lifecycle session and run its cursor, recording, configuration, and other cleanup hooks exactly once. Omit `session` to end the authenticated transport's implicit session. Idempotent.",
"platforms": [
"macos",
"windows",
"linux"
],
"capabilities": [
"session.lifecycle.end"
],
"annotations": {
"read_only": false,
"destructive": true,
"idempotent": true,
"open_world": false
},
"schema_mode": "canonical_runtime",
"cursor_semantics": {
"action": "system"
},
"input_schema": {
"additionalProperties": true,
"properties": {
"session": {
"description": "Optional public label to end. When omitted, end the caller's attached\nimplicit session.",
"type": "string"
}
},
"required": [],
"type": "object"
},
"success_output_schema": {
"additionalProperties": true,
"properties": {
"active": {
"const": false
},
"session": {
"type": "string"
}
},
"required": [
"session",
"active"
],
"type": "object"
}
},
{
"name": "escalate_session",
"description": "Deprecated compatibility tool for legacy capture-scope sessions. New callers select window or desktop modality on each action. No deescalate_session tool exists.",
"platforms": [
"macos",
"windows",
"linux"
],
"capabilities": [
"session.capture_scope.escalate"
],
"annotations": {
"read_only": false,
"destructive": false,
"idempotent": false,
"open_world": false
},
"schema_mode": "canonical_runtime",
"cursor_semantics": {
"action": "system"
},
"input_schema": {
"additionalProperties": true,
"properties": {
"detail": {
"description": "Optional bounded diagnostic detail. Never use secrets or page content.",
"maxLength": 200,
"type": "string"
},
"reason": {
"enum": [
"ax_tree_pixel_mismatch",
"background_delivery_failed",
"foreground_ineffective",
"no_window_target",
"other"
],
"type": "string"
},
"session": {
"type": "string"
}
},
"required": [
"session",
"reason"
],
"type": "object"
},
"success_output_schema": {
"additionalProperties": true,
"properties": {
"capture_scope": {
"enum": [
"auto",
"window",
"desktop"
],
"type": "string"
},
"desktop_unlocked": {
"type": "boolean"
},
"effective_scope": {
"enum": [
"window",
"desktop"
],
"type": "string"
},
"escalation_detail": {
"anyOf": [
{
"type": "string"
},
{
"type": "null"
}
]
},
"escalation_reason": {
"anyOf": [
{
"enum": [
"ax_tree_pixel_mismatch",
"background_delivery_failed",
"foreground_ineffective",
"no_window_target",
"other"
],
"type": "string"
},
{
"type": "null"
}
]
},
"session": {
"type": "string"
}
},
"required": [
"session",
"capture_scope",
"effective_scope",
"desktop_unlocked",
"escalation_reason",
"escalation_detail"
],
"type": "object"
}
},
{
"name": "get_agent_cursor_state",
"description": "Return the session cursor's theme, semantic playback, position, visibility, and motion.",
"platforms": [
"macos",
"windows",
"linux"
],
"capabilities": [
"agent_cursor.state"
],
"annotations": {
"read_only": true,
"destructive": false,
"idempotent": true,
"open_world": false
},
"schema_mode": "canonical_runtime",
"input_schema": {
"additionalProperties": false,
"properties": {
"session": {
"type": "string"
}
},
"required": [
"session"
],
"type": "object"
},
"success_output_schema": {
"additionalProperties": true,
"properties": {
"enabled": {
"type": "boolean"
},
"motion": {
"properties": {
"arc_flow": {
"type": "number"
},
"arc_size": {
"type": "number"
},
"dwell_after_click_ms": {
"type": "number"
},
"end_handle": {
"type": "number"
},
"glide_duration_ms": {
"type": "number"
},
"idle_hide_ms": {
"type": "number"
},
"spring": {
"type": "number"
},
"start_handle": {
"type": "number"
},
"turn_radius": {
"type": "number"
}
},
"required": [
"start_handle",
"end_handle",
"arc_size",
"arc_flow",
"spring",
"glide_duration_ms",
"dwell_after_click_ms",
"idle_hide_ms",
"turn_radius"
],
"type": "object"
},
"position": {
"properties": {
"x": {
"type": "number"
},
"y": {
"type": "number"
}
},
"required": [
"x",
"y"
],
"type": "object"
},
"session": {
"type": "string"
},
"theme": {
"properties": {
"fallback": {
"anyOf": [
{
"type": "string"
},
{
"type": "null"
}
]
},
"id": {
"type": "string"
},
"profile": {
"type": "string"
},
"reduced_motion": {
"enum": [
"auto",
"on",
"off"
],
"type": "string"
},
"version": {
"type": "string"
}
},
"required": [
"id",
"version",
"profile",
"reduced_motion",
"fallback"
],
"type": "object"
},
"visual_state": {
"properties": {
"frame": {
"minimum": 0,
"type": "integer"
},
"modifiers": {
"items": {
"type": "string"
},
"type": "array"
},
"phase": {
"type": "string"
},
"preempted_count": {
"minimum": 0,
"type": "integer"
},
"requested_action": {
"enum": [
"idle",
"observe",
"click",
"drag",
"scroll",
"text",
"key",
"navigate",
"app",
"transfer",
"record",
"system"
],
"type": "string"
},
"resolved_action": {
"enum": [
"idle",
"observe",
"click",
"drag",
"scroll",
"text",
"key",
"navigate",
"app",
"transfer",
"record",
"system"
],
"type": "string"
}
},
"required": [
"requested_action",
"resolved_action",
"modifiers",
"phase",
"frame",
"preempted_count"
],
"type": "object"
}
},
"required": [
"session",
"enabled",
"position",
"theme",
"visual_state",
"motion"
],
"type": "object"
}
},
{
"name": "get_cursor_position",
"description": "Return the OS cursor position when the platform can observe it.",
"platforms": [
"macos",
"windows",
"linux"
],
"capabilities": [
"screen.cursor.position"
],
"annotations": {
"read_only": true,
"destructive": false,
"idempotent": true,
"open_world": false
},
"schema_mode": "portable_subset",
"cursor_semantics": {
"action": "observe"
},
"input_schema": {
"additionalProperties": false,
"properties": {
"session": {
"description": "For multi-call work, prefer a short public session label and repeat it on every call that\naccepts it. Omit it to use the authenticated transport's implicit lifecycle session.",
"type": "string"
}
},
"required": [],
"type": "object"
},
"success_output_schema": {
"additionalProperties": true,
"properties": {
"available": {
"type": "boolean"
},
"source": {
"type": "string"
},
"x": {
"type": "number"
},
"y": {
"type": "number"
}
},
"type": "object"
}
},
{
"name": "get_desktop_state",
"description": "Capture the complete primary display at native resolution for a desktop-scope GUI loop.",
"platforms": [
"macos",
"windows",
"linux"
],
"capabilities": [
"screen.capture",
"screen.dimensions"
],
"annotations": {
"read_only": true,
"destructive": false,
"idempotent": false,
"open_world": false
},
"schema_mode": "portable_subset",
"cursor_semantics": {
"action": "observe"
},
"input_schema": {
"additionalProperties": false,
"properties": {
"screenshot_out_file": {
"description": "Write the PNG here instead of returning base64.",
"type": "string"
},
"session": {
"description": "For multi-call work, prefer a short public session label and repeat it on every call that\naccepts it. Omit it to use the authenticated transport's implicit lifecycle session.",
"type": "string"
}
},
"required": [],
"type": "object"
},
"success_output_schema": {
"additionalProperties": true,
"properties": {
"display": {
"type": "string"
},
"platform": {
"enum": [
"macos",
"linux",
"windows"
],
"type": "string"
},
"scale_factor": {
"type": "number"
},
"screen_height": {
"type": "integer"
},
"screen_width": {
"type": "integer"
},
"screenshot_file_path": {
"type": "string"
},
"screenshot_height": {
"type": "integer"
},
"screenshot_mime_type": {
"const": "image/png"
},
"screenshot_width": {
"type": "integer"
}
},
"required": [
"platform",
"display",
"screenshot_width",
"screenshot_height",
"screen_width",
"screen_height",
"scale_factor",
"screenshot_mime_type"
],
"type": "object"
}
},
{
"name": "get_screen_size",
"description": "Return the primary display dimensions and scale factor in the platform's desktop coordinate space.",
"platforms": [
"macos",
"windows",
"linux"
],
"capabilities": [
"screen.dimensions"
],
"annotations": {
"read_only": true,
"destructive": false,
"idempotent": true,
"open_world": false
},
"schema_mode": "portable_subset",
"cursor_semantics": {
"action": "observe"
},
"input_schema": {
"additionalProperties": false,
"properties": {
"session": {
"description": "For multi-call work, prefer a short public session label and repeat it on every call that\naccepts it. Omit it to use the authenticated transport's implicit lifecycle session.",
"type": "string"
}
},
"required": [],
"type": "object"
},
"success_output_schema": {
"additionalProperties": true,
"properties": {
"height": {
"type": "number"
},
"scale_factor": {
"type": "number"
},
"width": {
"type": "number"
}
},
"required": [
"width",
"height",
"scale_factor"
],
"type": "object"
}
},
{
"name": "get_session",
"description": "Read content-free lifecycle, cursor, recording, and idle status for one session visible to this authenticated transport. Omit `session` to inspect its implicit session.",
"platforms": [
"macos",
"windows",
"linux"
],
"capabilities": [
"session.lifecycle.read"
],
"annotations": {
"read_only": true,
"destructive": false,
"idempotent": true,
"open_world": false
},
"schema_mode": "canonical_runtime",
"cursor_semantics": {
"action": "system"
},
"input_schema": {
"additionalProperties": true,
"properties": {
"session": {
"description": "Optional public label. When omitted, inspect the caller's attached\nimplicit session.",
"type": "string"
}
},
"required": [],
"type": "object"
},
"success_output_schema": {
"additionalProperties": true,
"properties": {
"client_kind": {
"enum": [
"cli",
"direct",
"mcp",
"python_sdk",
"typescript_sdk"
],
"type": "string"
},
"cursor_visible": {
"type": "boolean"
},
"expires_in_seconds": {
"minimum": 0,
"type": "integer"
},
"idle_seconds": {
"minimum": 0,
"type": "integer"
},
"implicit": {
"type": "boolean"
},
"recording_active": {
"type": "boolean"
},
"session": {
"anyOf": [
{
"type": "string"
},
{
"type": "null"
}
]
},
"state": {
"enum": [
"active",
"ending"
],
"type": "string"
},
"transport": {
"enum": [
"cli",
"daemon",
"mcp_stdio",
"mcp_http"
],
"type": "string"
}
},
"required": [
"session",
"implicit",
"state",
"client_kind",
"transport",
"cursor_visible",
"recording_active",
"idle_seconds",
"expires_in_seconds"
],
"type": "object"
}
},
{
"name": "get_session_state",
"description": "Deprecated compatibility alias that reads a live legacy session's capture policy. Use get_session for lifecycle state.",
"platforms": [
"macos",
"windows",
"linux"
],
"capabilities": [
"session.capture_scope.read"
],
"annotations": {
"read_only": true,
"destructive": false,
"idempotent": true,
"open_world": false
},
"schema_mode": "canonical_runtime",
"cursor_semantics": {
"action": "system"
},
"input_schema": {
"additionalProperties": true,
"properties": {
"session": {
"description": "Optional public label. When omitted, inspect the caller's attached\nimplicit session.",
"type": "string"
}
},
"required": [],
"type": "object"
},
"success_output_schema": {
"additionalProperties": true,
"properties": {
"capture_scope": {
"enum": [
"auto",
"window",
"desktop"
],
"type": "string"
},
"desktop_unlocked": {
"type": "boolean"
},
"effective_scope": {
"enum": [
"window",
"desktop"
],
"type": "string"
},
"escalation_detail": {
"anyOf": [
{
"type": "string"
},
{
"type": "null"
}
]
},
"escalation_reason": {
"anyOf": [
{
"enum": [
"ax_tree_pixel_mismatch",
"background_delivery_failed",
"foreground_ineffective",
"no_window_target",
"other"
],
"type": "string"
},
{
"type": "null"
}
]
},
"session": {
"type": "string"
}
},
"required": [
"session",
"capture_scope",
"effective_scope",
"desktop_unlocked",
"escalation_reason",
"escalation_detail"
],
"type": "object"
}
},
{
"name": "get_window_state",
"description": "Walk one exact window's accessibility tree and return the actionable element snapshot, optionally as a versioned observation revision (`accessibility.observation_revision.v1`) diffed against a caller-supplied base.",
"platforms": [
"macos",
"windows",
"linux"
],
"capabilities": [
"accessibility.window_state",
"accessibility.tree",
"accessibility.tree.structured",
"accessibility.tree.bounded",
"accessibility.element_tokens",
"accessibility.observation_revision.v1",
"screen.capture",
"screen.capture.window"
],
"annotations": {
"read_only": true,
"destructive": false,
"idempotent": false,
"open_world": false
},
"schema_mode": "portable_subset",
"cursor_semantics": {
"action": "observe"
},
"input_schema": {
"additionalProperties": false,
"properties": {
"app_context": {
"description": "Use the compact macOS application observation with contextual menus.",
"type": "boolean"
},
"include_screenshot": {
"description": "Default true. Set false to skip the screenshot and return the tree only.",
"type": "boolean"
},
"max_depth": {
"description": "Cap on the accessibility-tree walk depth.",
"minimum": 1,
"type": "integer"
},
"max_elements": {
"description": "Cap on the total number of accessibility nodes walked.",
"minimum": 1,
"type": "integer"
},
"observation_revision": {
"additionalProperties": false,
"description": "Opt in to `accessibility.observation_revision.v1`. Requires a bound\ndriver session. Omit to preserve the legacy full-snapshot contract.",
"properties": {
"base_revision_id": {
"maxLength": 256,
"minLength": 1,
"type": "string"
},
"force_full": {
"default": false,
"type": "boolean"
},
"projection_version": {
"maxLength": 128,
"minLength": 1,
"type": "string"
},
"serializer_version": {
"maxLength": 128,
"minLength": 1,
"type": "string"
},
"version": {
"maximum": 1,
"minimum": 1,
"type": "integer"
}
},
"required": [
"version",
"serializer_version",
"projection_version"
],
"type": "object"
},
"pid": {
"description": "Target process ID.",
"minimum": 1,
"type": "integer"
},
"query": {
"description": "Case-insensitive projection filter. Incompatible with `observation_revision`.",
"type": "string"
},
"screenshot_out_file": {
"description": "Write the PNG here instead of returning base64.",
"type": "string"
},
"session": {
"description": "For multi-call work, prefer a short public session label and repeat it on every call that\naccepts it. Omit it to use the authenticated transport's implicit lifecycle session.",
"type": "string"
},
"window_id": {
"description": "Exact window to observe.",
"minimum": 1,
"type": "integer"
}
},
"required": [
"pid",
"window_id"
],
"type": "object"
},
"success_output_schema": {
"additionalProperties": true,
"properties": {
"observation_revision": {
"additionalProperties": true,
"properties": {
"base_revision_id": {
"type": "string"
},
"capability": {
"const": "accessibility.observation_revision.v1"
},
"lineage_id": {
"minLength": 1,
"type": "string"
},
"mode": {
"enum": [
"full",
"diff",
"no_change"
],
"type": "string"
},
"projection_version": {
"minLength": 1,
"type": "string"
},
"resync_reason": {
"type": "string"
},
"revision_id": {
"minLength": 1,
"type": "string"
},
"serializer_version": {
"minLength": 1,
"type": "string"
},
"stable_element_ids": {
"type": [
"boolean",
"null"
]
},
"version": {
"type": "integer"
}
},
"required": [
"capability",
"version",
"serializer_version",
"projection_version",
"mode",
"lineage_id",
"revision_id"
],
"type": [
"object",
"null"
]
}
},
"type": "object"
}
},
{
"name": "hotkey",
"description": "Press a key chord in the foreground desktop application.",
"platforms": [
"macos",
"windows",
"linux"
],
"capabilities": [
"input.keyboard.hotkey",
"accessibility.element_tokens"
],
"annotations": {
"read_only": false,
"destructive": true,
"idempotent": false,
"open_world": true
},
"schema_mode": "portable_subset",
"cursor_semantics": {
"action": "key"
},
"input_schema": {
"additionalProperties": false,
"properties": {
"keys": {
"items": {
"type": "string"
},
"minItems": 2,
"type": "array"
},
"scope": {
"description": "Deprecated flat desktop target retained for wire compatibility.",
"enum": [
"desktop"
],
"type": "string"
},
"session": {
"description": "For multi-call work, prefer a short public session label and repeat it on every call that\naccepts it. Omit it to use the authenticated transport's implicit lifecycle session.",
"type": "string"
},
"target": {
"additionalProperties": true,
"description": "Exact window or desktop target. Window requires pid and window_id; desktop requires display_id.",
"properties": {
"display_id": {
"description": "Use primary for the primary display.",
"type": "string"
},
"kind": {
"enum": [
"window",
"desktop"
],
"type": "string"
},
"pid": {
"minimum": 0,
"type": "integer"
},
"window_id": {
"minimum": 0,
"type": "integer"
}
},
"required": [
"kind"
],
"type": "object"
}
},
"required": [
"keys"
],
"type": "object"
},
"success_output_schema": {
"additionalProperties": false,
"properties": {
"delivery": {
"additionalProperties": false,
"properties": {
"delivered_count": {
"minimum": 0,
"type": [
"integer",
"null"
]
},
"mode": {
"enum": [
"background",
"foreground",
"not_applicable",
"unknown"
],
"type": "string"
}
},
"required": [
"mode"
],
"type": [
"object",
"null"
]
},
"effect": {
"enum": [
"confirmed",
"partial",
"unverifiable",
"suspected_noop",
"refused"
],
"type": "string"
},
"escalation": {
"additionalProperties": false,
"properties": {
"reason": {
"enum": [
"route_unavailable",
"delivery_failed",
"effect_unconfirmed",
"suspected_noop",
"permission_required"
],
"type": "string"
},
"target": {
"enum": [
"pixel",
"foreground",
"page",
"session"
],
"type": "string"
}
},
"required": [
"target",
"reason"
],
"type": [
"object",
"null"
]
},
"evidence": {
"items": {
"additionalProperties": false,
"properties": {
"kind": {
"enum": [
"value_readback",
"window_change"
],
"type": "string"
}
},
"required": [
"kind"
],
"type": "object"
},
"type": [
"array",
"null"
]
},
"route": {
"enum": [
"accessibility",
"synthetic_events",
"global_input",
"system_api",
"dom",
"trusted_input"
],
"type": "string"
}
},
"required": [
"effect",
"route"
],
"type": "object"
}
},
{
"name": "invoke_menu",
"description": "Resolve an exact application-menu path one live native level at a time and invoke its final item through accessibility APIs. Missing, ambiguous, disabled, or structurally mismatched segments fail closed; this tool never falls back to pixels.",
"platforms": [
"macos",
"windows",
"linux"
],
"capabilities": [
"menu.path.invoke",
"accessibility.menu.native"
],
"annotations": {
"read_only": false,
"destructive": true,
"idempotent": false,
"open_world": true
},
"schema_mode": "portable_subset",
"cursor_semantics": {
"action": "app"
},
"input_schema": {
"additionalProperties": false,
"description": "Exact, immediate-child application menu path to resolve and invoke through\nthe operating system's accessibility API. Path labels are matched after\ntrimming surrounding whitespace and otherwise remain case-sensitive.",
"properties": {
"path": {
"items": {
"maxLength": 200,
"minLength": 1,
"type": "string"
},
"maxItems": 16,
"minItems": 1,
"type": "array"
},
"pid": {
"minimum": 1,
"type": "integer"
},
"session": {
"description": "For multi-call work, prefer a short public session label and repeat it on every call that\naccepts it. Omit it to use the authenticated transport's implicit lifecycle session.",
"type": "string"
},
"window_id": {
"minimum": 1,
"type": "integer"
}
},
"required": [
"pid",
"window_id",
"path"
],
"type": "object"
},
"success_output_schema": {
"additionalProperties": false,
"properties": {
"delivery": {
"additionalProperties": false,
"properties": {
"delivered_count": {
"minimum": 0,
"type": [
"integer",
"null"
]
},
"mode": {
"enum": [
"background",
"foreground",
"not_applicable",
"unknown"
],
"type": "string"
}
},
"required": [
"mode"
],
"type": [
"object",
"null"
]
},
"effect": {
"enum": [
"confirmed",
"partial",
"unverifiable",
"suspected_noop",
"refused"
],
"type": "string"
},
"escalation": {
"additionalProperties": false,
"properties": {
"reason": {
"enum": [
"route_unavailable",
"delivery_failed",
"effect_unconfirmed",
"suspected_noop",
"permission_required"
],
"type": "string"
},
"target": {
"enum": [
"pixel",
"foreground",
"page",
"session"
],
"type": "string"
}
},
"required": [
"target",
"reason"
],
"type": [
"object",
"null"
]
},
"evidence": {
"items": {
"additionalProperties": false,
"properties": {
"kind": {
"enum": [
"value_readback",
"window_change"
],
"type": "string"
}
},
"required": [
"kind"
],
"type": "object"
},
"type": [
"array",
"null"
]
},
"route": {
"enum": [
"accessibility",
"synthetic_events",
"global_input",
"system_api",
"dom",
"trusted_input"
],
"type": "string"
}
},
"required": [
"effect",
"route"
],
"type": "object"
}
},
{
"name": "list_sessions",
"description": "List content-free lifecycle summaries attached to this authenticated transport lease. It does not enumerate other callers' sessions.",
"platforms": [
"macos",
"windows",
"linux"
],
"capabilities": [
"session.lifecycle.list"
],
"annotations": {
"read_only": true,
"destructive": false,
"idempotent": true,
"open_world": false
},
"schema_mode": "canonical_runtime",
"cursor_semantics": {
"action": "system"
},
"input_schema": {
"additionalProperties": true,
"properties": {
"cursor": {
"description": "Opaque continuation cursor returned by a previous call.",
"type": "string"
},
"limit": {
"description": "Maximum number of content-free summaries to return (default 50, max\n100). Ordinary agent transports are scoped to their own lease.",
"minimum": 0,
"type": "integer"
}
},
"required": [],
"type": "object"
},
"success_output_schema": {
"additionalProperties": true,
"properties": {
"next_cursor": {
"anyOf": [
{
"type": "string"
},
{
"type": "null"
}
]
},
"sessions": {
"items": {
"properties": {
"client_kind": {
"enum": [
"cli",
"direct",
"mcp",
"python_sdk",
"typescript_sdk"
],
"type": "string"
},
"cursor_visible": {
"type": "boolean"
},
"expires_in_seconds": {
"minimum": 0,
"type": "integer"
},
"idle_seconds": {
"minimum": 0,
"type": "integer"
},
"implicit": {
"type": "boolean"
},
"recording_active": {
"type": "boolean"
},
"session": {
"anyOf": [
{
"type": "string"
},
{
"type": "null"
}
]
},
"state": {
"enum": [
"active",
"ending"
],
"type": "string"
},
"transport": {
"enum": [
"cli",
"daemon",
"mcp_stdio",
"mcp_http"
],
"type": "string"
}
},
"required": [
"session",
"implicit",
"state",
"client_kind",
"transport",
"cursor_visible",
"recording_active",
"idle_seconds",
"expires_in_seconds"
],
"type": "object"
},
"type": "array"
}
},
"required": [
"sessions",
"next_cursor"
],
"type": "object"
}
},
{
"name": "move_cursor",
"description": "Move the real OS pointer in get_desktop_state coordinates.",
"platforms": [
"macos",
"windows",
"linux"
],
"capabilities": [
"agent_cursor.move",
"input.pointer.move"
],
"annotations": {
"read_only": false,
"destructive": false,
"idempotent": true,
"open_world": false
},
"schema_mode": "portable_subset",
"cursor_semantics": {
"action": "navigate"
},
"input_schema": {
"additionalProperties": false,
"properties": {
"scope": {
"description": "Deprecated flat desktop target retained for wire compatibility.",
"enum": [
"desktop"
],
"type": "string"
},
"session": {
"description": "For multi-call work, prefer a short public session label and repeat it on every call that\naccepts it. Omit it to use the authenticated transport's implicit lifecycle session.",
"type": "string"
},
"target": {
"additionalProperties": true,
"description": "Preferred per-call target. New callers should set this field.",
"properties": {
"display_id": {
"description": "Use primary for the primary display.",
"type": "string"
},
"kind": {
"enum": [
"window",
"desktop"
],
"type": "string"
},
"pid": {
"minimum": 0,
"type": "integer"
},
"window_id": {
"minimum": 0,
"type": "integer"
}
},
"required": [
"kind"
],
"type": "object"
},
"x": {
"type": "number"
},
"y": {
"type": "number"
}
},
"required": [
"x",
"y"
],
"type": "object"
},
"success_output_schema": {
"additionalProperties": false,
"properties": {
"delivery": {
"additionalProperties": false,
"properties": {
"delivered_count": {
"minimum": 0,
"type": [
"integer",
"null"
]
},
"mode": {
"enum": [
"background",
"foreground",
"not_applicable",
"unknown"
],
"type": "string"
}
},
"required": [
"mode"
],
"type": [
"object",
"null"
]
},
"effect": {
"enum": [
"confirmed",
"partial",
"unverifiable",
"suspected_noop",
"refused"
],
"type": "string"
},
"escalation": {
"additionalProperties": false,
"properties": {
"reason": {
"enum": [
"route_unavailable",
"delivery_failed",
"effect_unconfirmed",
"suspected_noop",
"permission_required"
],
"type": "string"
},
"target": {
"enum": [
"pixel",
"foreground",
"page",
"session"
],
"type": "string"
}
},
"required": [
"target",
"reason"
],
"type": [
"object",
"null"
]
},
"evidence": {
"items": {
"additionalProperties": false,
"properties": {
"kind": {
"enum": [
"value_readback",
"window_change"
],
"type": "string"
}
},
"required": [
"kind"
],
"type": "object"
},
"type": [
"array",
"null"
]
},
"route": {
"enum": [
"accessibility",
"synthetic_events",
"global_input",
"system_api",
"dom",
"trusted_input"
],
"type": "string"
}
},
"required": [
"effect",
"route"
],
"type": "object"
}
},
{
"name": "paste",
"description": "Paste plain text, Markdown, or HTML into the exact app window, then restore the previous clipboard if no external change replaced it. macOS only.",
"platforms": [
"macos"
],
"capabilities": [
"input.keyboard.paste",
"clipboard.read",
"clipboard.write"
],
"annotations": {
"read_only": false,
"destructive": true,
"idempotent": false,
"open_world": true
},
"schema_mode": "canonical_runtime",
"cursor_semantics": {
"action": "text"
},
"input_schema": {
"additionalProperties": false,
"properties": {
"app_context": {
"type": "boolean"
},
"format": {
"default": "text",
"enum": [
"text",
"md",
"html"
],
"type": "string"
},
"pid": {
"minimum": 1,
"type": "integer"
},
"text": {
"type": "string"
},
"window_id": {
"minimum": 1,
"type": "integer"
}
},
"required": [
"pid",
"window_id",
"text"
],
"type": "object"
},
"success_output_schema": {
"additionalProperties": false,
"properties": {
"delivery": {
"additionalProperties": false,
"properties": {
"delivered_count": {
"minimum": 0,
"type": [
"integer",
"null"
]
},
"mode": {
"enum": [
"background",
"foreground",
"not_applicable",
"unknown"
],
"type": "string"
}
},
"required": [
"mode"
],
"type": [
"object",
"null"
]
},
"effect": {
"enum": [
"confirmed",
"partial",
"unverifiable",
"suspected_noop",
"refused"
],
"type": "string"
},
"escalation": {
"additionalProperties": false,
"properties": {
"reason": {
"enum": [
"route_unavailable",
"delivery_failed",
"effect_unconfirmed",
"suspected_noop",
"permission_required"
],
"type": "string"
},
"target": {
"enum": [
"pixel",
"foreground",
"page",
"session"
],
"type": "string"
}
},
"required": [
"target",
"reason"
],
"type": [
"object",
"null"
]
},
"evidence": {
"items": {
"additionalProperties": false,
"properties": {
"kind": {
"enum": [
"value_readback",
"window_change"
],
"type": "string"
}
},
"required": [
"kind"
],
"type": "object"
},
"type": [
"array",
"null"
]
},
"route": {
"enum": [
"accessibility",
"synthetic_events",
"global_input",
"system_api",
"dom",
"trusted_input"
],
"type": "string"
}
},
"required": [
"effect",
"route"
],
"type": "object"
}
},
{
"name": "perform_secondary_action",
"description": "Perform one explicitly named action advertised by the exact current accessibility element. Missing, ambiguous, disabled, or stale actions fail closed without a pixel fallback.",
"platforms": [
"macos",
"windows",
"linux"
],
"capabilities": [
"accessibility.action.secondary",
"accessibility.element_tokens"
],
"annotations": {
"read_only": false,
"destructive": true,
"idempotent": false,
"open_world": true
},
"schema_mode": "portable_subset",
"cursor_semantics": {
"action": "app"
},
"input_schema": {
"additionalProperties": false,
"properties": {
"action": {
"type": "string"
},
"element_token": {
"maxLength": 512,
"minLength": 1,
"type": "string"
},
"pid": {
"minimum": 1,
"type": "integer"
},
"window_id": {
"minimum": 1,
"type": "integer"
}
},
"required": [
"pid",
"element_token",
"action"
],
"type": "object"
},
"success_output_schema": {
"additionalProperties": false,
"properties": {
"delivery": {
"additionalProperties": false,
"properties": {
"delivered_count": {
"minimum": 0,
"type": [
"integer",
"null"
]
},
"mode": {
"enum": [
"background",
"foreground",
"not_applicable",
"unknown"
],
"type": "string"
}
},
"required": [
"mode"
],
"type": [
"object",
"null"
]
},
"effect": {
"enum": [
"confirmed",
"partial",
"unverifiable",
"suspected_noop",
"refused"
],
"type": "string"
},
"escalation": {
"additionalProperties": false,
"properties": {
"reason": {
"enum": [
"route_unavailable",
"delivery_failed",
"effect_unconfirmed",
"suspected_noop",
"permission_required"
],
"type": "string"
},
"target": {
"enum": [
"pixel",
"foreground",
"page",
"session"
],
"type": "string"
}
},
"required": [
"target",
"reason"
],
"type": [
"object",
"null"
]
},
"evidence": {
"items": {
"additionalProperties": false,
"properties": {
"kind": {
"enum": [
"value_readback",
"window_change"
],
"type": "string"
}
},
"required": [
"kind"
],
"type": "object"
},
"type": [
"array",
"null"
]
},
"route": {
"enum": [
"accessibility",
"synthetic_events",
"global_input",
"system_api",
"dom",
"trusted_input"
],
"type": "string"
}
},
"required": [
"effect",
"route"
],
"type": "object"
}
},
{
"name": "press_key",
"description": "Press one key, with optional modifiers, in the foreground desktop application.",
"platforms": [
"macos",
"windows",
"linux"
],
"capabilities": [
"input.keyboard.press",
"accessibility.element_tokens"
],
"annotations": {
"read_only": false,
"destructive": true,
"idempotent": false,
"open_world": true
},
"schema_mode": "portable_subset",
"cursor_semantics": {
"action": "key"
},
"input_schema": {
"additionalProperties": false,
"properties": {
"key": {
"type": "string"
},
"modifiers": {
"items": {
"type": "string"
},
"type": "array"
},
"scope": {
"description": "Deprecated flat desktop target retained for wire compatibility.",
"enum": [
"desktop"
],
"type": "string"
},
"session": {
"description": "For multi-call work, prefer a short public session label and repeat it on every call that\naccepts it. Omit it to use the authenticated transport's implicit lifecycle session.",
"type": "string"
},
"target": {
"additionalProperties": true,
"description": "Exact window or desktop target. Window requires pid and window_id; desktop requires display_id.",
"properties": {
"display_id": {
"description": "Use primary for the primary display.",
"type": "string"
},
"kind": {
"enum": [
"window",
"desktop"
],
"type": "string"
},
"pid": {
"minimum": 0,
"type": "integer"
},
"window_id": {
"minimum": 0,
"type": "integer"
}
},
"required": [
"kind"
],
"type": "object"
}
},
"required": [
"key"
],
"type": "object"
},
"success_output_schema": {
"additionalProperties": false,
"properties": {
"delivery": {
"additionalProperties": false,
"properties": {
"delivered_count": {
"minimum": 0,
"type": [
"integer",
"null"
]
},
"mode": {
"enum": [
"background",
"foreground",
"not_applicable",
"unknown"
],
"type": "string"
}
},
"required": [
"mode"
],
"type": [
"object",
"null"
]
},
"effect": {
"enum": [
"confirmed",
"partial",
"unverifiable",
"suspected_noop",
"refused"
],
"type": "string"
},
"escalation": {
"additionalProperties": false,
"properties": {
"reason": {
"enum": [
"route_unavailable",
"delivery_failed",
"effect_unconfirmed",
"suspected_noop",
"permission_required"
],
"type": "string"
},
"target": {
"enum": [
"pixel",
"foreground",
"page",
"session"
],
"type": "string"
}
},
"required": [
"target",
"reason"
],
"type": [
"object",
"null"
]
},
"evidence": {
"items": {
"additionalProperties": false,
"properties": {
"kind": {
"enum": [
"value_readback",
"window_change"
],
"type": "string"
}
},
"required": [
"kind"
],
"type": "object"
},
"type": [
"array",
"null"
]
},
"route": {
"enum": [
"accessibility",
"synthetic_events",
"global_input",
"system_api",
"dom",
"trusted_input"
],
"type": "string"
}
},
"required": [
"effect",
"route"
],
"type": "object"
}
},
{
"name": "scroll",
"description": "Scroll at an absolute point in get_desktop_state coordinates.",
"platforms": [
"macos",
"windows",
"linux"
],
"capabilities": [
"input.pointer.scroll",
"accessibility.element_tokens"
],
"annotations": {
"read_only": false,
"destructive": false,
"idempotent": false,
"open_world": true
},
"schema_mode": "portable_subset",
"cursor_semantics": {
"action": "scroll"
},
"input_schema": {
"additionalProperties": false,
"properties": {
"amount": {
"maximum": 50,
"minimum": 1,
"type": "integer"
},
"by": {
"enum": [
"line",
"page"
],
"type": "string"
},
"direction": {
"enum": [
"up",
"down",
"left",
"right"
],
"type": "string"
},
"scope": {
"description": "Deprecated flat desktop target retained for wire compatibility.",
"enum": [
"desktop"
],
"type": "string"
},
"session": {
"description": "For multi-call work, prefer a short public session label and repeat it on every call that\naccepts it. Omit it to use the authenticated transport's implicit lifecycle session.",
"type": "string"
},
"target": {
"additionalProperties": true,
"description": "Exact window or desktop target. Window requires pid and window_id; desktop requires display_id.",
"properties": {
"display_id": {
"description": "Use primary for the primary display.",
"type": "string"
},
"kind": {
"enum": [
"window",
"desktop"
],
"type": "string"
},
"pid": {
"minimum": 0,
"type": "integer"
},
"window_id": {
"minimum": 0,
"type": "integer"
}
},
"required": [
"kind"
],
"type": "object"
},
"x": {
"type": "number"
},
"y": {
"type": "number"
}
},
"required": [
"x",
"y",
"direction"
],
"type": "object"
},
"success_output_schema": {
"additionalProperties": false,
"properties": {
"delivery": {
"additionalProperties": false,
"properties": {
"delivered_count": {
"minimum": 0,
"type": [
"integer",
"null"
]
},
"mode": {
"enum": [
"background",
"foreground",
"not_applicable",
"unknown"
],
"type": "string"
}
},
"required": [
"mode"
],
"type": [
"object",
"null"
]
},
"effect": {
"enum": [
"confirmed",
"partial",
"unverifiable",
"suspected_noop",
"refused"
],
"type": "string"
},
"escalation": {
"additionalProperties": false,
"properties": {
"reason": {
"enum": [
"route_unavailable",
"delivery_failed",
"effect_unconfirmed",
"suspected_noop",
"permission_required"
],
"type": "string"
},
"target": {
"enum": [
"pixel",
"foreground",
"page",
"session"
],
"type": "string"
}
},
"required": [
"target",
"reason"
],
"type": [
"object",
"null"
]
},
"evidence": {
"items": {
"additionalProperties": false,
"properties": {
"kind": {
"enum": [
"value_readback",
"window_change"
],
"type": "string"
}
},
"required": [
"kind"
],
"type": "object"
},
"type": [
"array",
"null"
]
},
"route": {
"enum": [
"accessibility",
"synthetic_events",
"global_input",
"system_api",
"dom",
"trusted_input"
],
"type": "string"
}
},
"required": [
"effect",
"route"
],
"type": "object"
}
},
{
"name": "select_text",
"description": "Select one unique text match in the exact accessibility element, optionally constrained by adjacent prefix/suffix, or place its caret before/after the match. macOS only.",
"platforms": [
"macos"
],
"capabilities": [
"accessibility.text.select",
"accessibility.element_tokens"
],
"annotations": {
"read_only": false,
"destructive": true,
"idempotent": true,
"open_world": true
},
"schema_mode": "canonical_runtime",
"cursor_semantics": {
"action": "text"
},
"input_schema": {
"additionalProperties": false,
"properties": {
"element_token": {
"maxLength": 512,
"minLength": 1,
"type": "string"
},
"pid": {
"minimum": 1,
"type": "integer"
},
"prefix": {
"type": "string"
},
"selection": {
"default": "text",
"enum": [
"text",
"cursor_before",
"cursor_after"
],
"type": "string"
},
"suffix": {
"type": "string"
},
"text": {
"type": "string"
},
"window_id": {
"minimum": 1,
"type": "integer"
}
},
"required": [
"pid",
"window_id",
"element_token",
"text"
],
"type": "object"
},
"success_output_schema": {
"additionalProperties": false,
"properties": {
"delivery": {
"additionalProperties": false,
"properties": {
"delivered_count": {
"minimum": 0,
"type": [
"integer",
"null"
]
},
"mode": {
"enum": [
"background",
"foreground",
"not_applicable",
"unknown"
],
"type": "string"
}
},
"required": [
"mode"
],
"type": [
"object",
"null"
]
},
"effect": {
"enum": [
"confirmed",
"partial",
"unverifiable",
"suspected_noop",
"refused"
],
"type": "string"
},
"escalation": {
"additionalProperties": false,
"properties": {
"reason": {
"enum": [
"route_unavailable",
"delivery_failed",
"effect_unconfirmed",
"suspected_noop",
"permission_required"
],
"type": "string"
},
"target": {
"enum": [
"pixel",
"foreground",
"page",
"session"
],
"type": "string"
}
},
"required": [
"target",
"reason"
],
"type": [
"object",
"null"
]
},
"evidence": {
"items": {
"additionalProperties": false,
"properties": {
"kind": {
"enum": [
"value_readback",
"window_change"
],
"type": "string"
}
},
"required": [
"kind"
],
"type": "object"
},
"type": [
"array",
"null"
]
},
"route": {
"enum": [
"accessibility",
"synthetic_events",
"global_input",
"system_api",
"dom",
"trusted_input"
],
"type": "string"
}
},
"required": [
"effect",
"route"
],
"type": "object"
}
},
{
"name": "set_agent_cursor_enabled",
"description": "Show or hide the agent cursor owned by a session.",
"platforms": [
"macos",
"windows",
"linux"
],
"capabilities": [
"agent_cursor.set_enabled"
],
"annotations": {
"read_only": false,
"destructive": false,
"idempotent": true,
"open_world": false
},
"schema_mode": "canonical_runtime",
"input_schema": {
"additionalProperties": false,
"properties": {
"enabled": {
"type": "boolean"
},
"session": {
"type": "string"
}
},
"required": [
"session",
"enabled"
],
"type": "object"
},
"success_output_schema": {
"additionalProperties": true,
"properties": {
"enabled": {
"type": "boolean"
},
"session": {
"type": "string"
}
},
"required": [
"session",
"enabled"
],
"type": "object"
}
},
{
"name": "set_agent_cursor_motion",
"description": "Configure only movement physics and visibility timing for a session cursor.",
"platforms": [
"macos",
"windows",
"linux"
],
"capabilities": [
"agent_cursor.set_motion"
],
"annotations": {
"read_only": false,
"destructive": false,
"idempotent": true,
"open_world": false
},
"schema_mode": "canonical_runtime",
"input_schema": {
"additionalProperties": false,
"properties": {
"arc_flow": {
"type": "number"
},
"arc_size": {
"type": "number"
},
"dwell_after_click_ms": {
"type": "number"
},
"end_handle": {
"type": "number"
},
"glide_duration_ms": {
"type": "number"
},
"idle_hide_ms": {
"type": "number"
},
"session": {
"type": "string"
},
"spring": {
"type": "number"
},
"start_handle": {
"type": "number"
},
"turn_radius": {
"type": "number"
}
},
"required": [
"session"
],
"type": "object"
},
"success_output_schema": {
"additionalProperties": true,
"properties": {
"motion": {
"properties": {
"arc_flow": {
"type": "number"
},
"arc_size": {
"type": "number"
},
"dwell_after_click_ms": {
"type": "number"
},
"end_handle": {
"type": "number"
},
"glide_duration_ms": {
"type": "number"
},
"idle_hide_ms": {
"type": "number"
},
"spring": {
"type": "number"
},
"start_handle": {
"type": "number"
},
"turn_radius": {
"type": "number"
}
},
"required": [
"start_handle",
"end_handle",
"arc_size",
"arc_flow",
"spring",
"glide_duration_ms",
"dwell_after_click_ms",
"idle_hide_ms",
"turn_radius"
],
"type": "object"
},
"session": {
"type": "string"
}
},
"required": [
"session",
"motion"
],
"type": "object"
}
},
{
"name": "set_agent_cursor_theme",
"description": "Select an already-installed cursor theme for a session.",
"platforms": [
"macos",
"windows",
"linux"
],
"capabilities": [
"agent_cursor.set_theme"
],
"annotations": {
"read_only": false,
"destructive": false,
"idempotent": true,
"open_world": false
},
"schema_mode": "canonical_runtime",
"input_schema": {
"additionalProperties": false,
"properties": {
"reduced_motion": {
"default": "auto",
"enum": [
"auto",
"on",
"off"
],
"type": "string"
},
"session": {
"type": "string"
},
"theme_id": {
"maxLength": 200,
"minLength": 1,
"type": "string"
}
},
"required": [
"session",
"theme_id"
],
"type": "object"
},
"success_output_schema": {
"additionalProperties": true,
"properties": {
"session": {
"type": "string"
},
"theme": {
"properties": {
"fallback": {
"anyOf": [
{
"type": "string"
},
{
"type": "null"
}
]
},
"id": {
"type": "string"
},
"profile": {
"type": "string"
},
"reduced_motion": {
"enum": [
"auto",
"on",
"off"
],
"type": "string"
},
"version": {
"type": "string"
}
},
"required": [
"id",
"version",
"profile",
"reduced_motion",
"fallback"
],
"type": "object"
}
},
"required": [
"session",
"theme"
],
"type": "object"
}
},
{
"name": "set_value",
"description": "Set the value of the exact current accessibility element named by an opaque element token.",
"platforms": [
"macos",
"windows",
"linux"
],
"capabilities": [
"accessibility.value.set",
"accessibility.element_tokens"
],
"annotations": {
"read_only": false,
"destructive": true,
"idempotent": true,
"open_world": true
},
"schema_mode": "portable_subset",
"cursor_semantics": {
"action": "text"
},
"input_schema": {
"additionalProperties": false,
"properties": {
"element_token": {
"maxLength": 512,
"minLength": 1,
"type": "string"
},
"pid": {
"minimum": 1,
"type": "integer"
},
"value": {
"type": "string"
},
"window_id": {
"minimum": 1,
"type": "integer"
}
},
"required": [
"pid",
"element_token",
"value"
],
"type": "object"
},
"success_output_schema": {
"additionalProperties": false,
"properties": {
"delivery": {
"additionalProperties": false,
"properties": {
"delivered_count": {
"minimum": 0,
"type": [
"integer",
"null"
]
},
"mode": {
"enum": [
"background",
"foreground",
"not_applicable",
"unknown"
],
"type": "string"
}
},
"required": [
"mode"
],
"type": [
"object",
"null"
]
},
"effect": {
"enum": [
"confirmed",
"partial",
"unverifiable",
"suspected_noop",
"refused"
],
"type": "string"
},
"escalation": {
"additionalProperties": false,
"properties": {
"reason": {
"enum": [
"route_unavailable",
"delivery_failed",
"effect_unconfirmed",
"suspected_noop",
"permission_required"
],
"type": "string"
},
"target": {
"enum": [
"pixel",
"foreground",
"page",
"session"
],
"type": "string"
}
},
"required": [
"target",
"reason"
],
"type": [
"object",
"null"
]
},
"evidence": {
"items": {
"additionalProperties": false,
"properties": {
"kind": {
"enum": [
"value_readback",
"window_change"
],
"type": "string"
}
},
"required": [
"kind"
],
"type": "object"
},
"type": [
"array",
"null"
]
},
"route": {
"enum": [
"accessibility",
"synthetic_events",
"global_input",
"system_api",
"dom",
"trusted_input"
],
"type": "string"
}
},
"required": [
"effect",
"route"
],
"type": "object"
}
},
{
"name": "set_window_frame",
"description": "Set one exact top-level window's frame in the desktop-coordinate space reported by list_windows and verify the resulting geometry through an independent readback.",
"platforms": [
"macos",
"windows",
"linux"
],
"capabilities": [
"window.frame.set"
],
"annotations": {
"read_only": false,
"destructive": false,
"idempotent": true,
"open_world": false
},
"schema_mode": "portable_subset",
"cursor_semantics": {
"action": "app"
},
"input_schema": {
"additionalProperties": false,
"properties": {
"height": {
"minimum": 1,
"type": "number"
},
"pid": {
"minimum": 1,
"type": "integer"
},
"session": {
"description": "For multi-call work, prefer a short public session label and repeat it on every call that\naccepts it. Omit it to use the authenticated transport's implicit lifecycle session.",
"type": "string"
},
"width": {
"minimum": 1,
"type": "number"
},
"window_id": {
"minimum": 1,
"type": "integer"
},
"x": {
"type": "number"
},
"y": {
"type": "number"
}
},
"required": [
"pid",
"window_id",
"x",
"y",
"width",
"height"
],
"type": "object"
},
"success_output_schema": {
"additionalProperties": false,
"properties": {
"delivery": {
"additionalProperties": false,
"properties": {
"delivered_count": {
"minimum": 0,
"type": [
"integer",
"null"
]
},
"mode": {
"enum": [
"background",
"foreground",
"not_applicable",
"unknown"
],
"type": "string"
}
},
"required": [
"mode"
],
"type": [
"object",
"null"
]
},
"effect": {
"enum": [
"confirmed",
"partial",
"unverifiable",
"suspected_noop",
"refused"
],
"type": "string"
},
"escalation": {
"additionalProperties": false,
"properties": {
"reason": {
"enum": [
"route_unavailable",
"delivery_failed",
"effect_unconfirmed",
"suspected_noop",
"permission_required"
],
"type": "string"
},
"target": {
"enum": [
"pixel",
"foreground",
"page",
"session"
],
"type": "string"
}
},
"required": [
"target",
"reason"
],
"type": [
"object",
"null"
]
},
"evidence": {
"items": {
"additionalProperties": false,
"properties": {
"kind": {
"enum": [
"value_readback",
"window_change"
],
"type": "string"
}
},
"required": [
"kind"
],
"type": "object"
},
"type": [
"array",
"null"
]
},
"route": {
"enum": [
"accessibility",
"synthetic_events",
"global_input",
"system_api",
"dom",
"trusted_input"
],
"type": "string"
}
},
"required": [
"effect",
"route"
],
"type": "object"
}
},
{
"name": "start_session",
"description": "Optionally create or return a lifecycle session before acting. For multi-call work, prefer a short public `session` label and repeat it on every call that accepts it; an omitted value uses the authenticated transport lease's implicit session instead. This tool is optional because an ordinary action can create or reuse a named run directly. Use it to set the initial cursor theme before acting or to revive a public name after it has ended; ordinary actions never revive ended names. `capture_scope` is deprecated compatibility input; new callers select window or desktop modality per action. Idempotent.",
"platforms": [
"macos",
"windows",
"linux"
],
"capabilities": [
"session.lifecycle.start",
"session.capture_scope"
],
"annotations": {
"read_only": false,
"destructive": false,
"idempotent": true,
"open_world": false
},
"schema_mode": "canonical_runtime",
"cursor_semantics": {
"action": "system"
},
"input_schema": {
"additionalProperties": true,
"properties": {
"capture_scope": {
"description": "Deprecated compatibility policy. New callers select window or desktop\nmodality on each action instead of storing it on the session.",
"enum": [
"auto",
"window",
"desktop"
],
"type": "string"
},
"cursor_theme": {
"additionalProperties": true,
"description": "Optional initial cursor theme. The host applies it before the cursor is\nfirst made visible, avoiding a flash of the default theme.",
"properties": {
"reduced_motion": {
"default": "auto",
"enum": [
"auto",
"on",
"off"
],
"type": "string"
},
"theme_id": {
"type": "string"
}
},
"required": [
"theme_id"
],
"type": "object"
},
"session": {
"description": "Optional stable public label for this run (e.g. \"research-run-1\").\nWhen omitted, the authenticated transport lease's implicit session is\ncreated or returned.",
"type": "string"
}
},
"required": [],
"type": "object"
},
"success_output_schema": {
"additionalProperties": true,
"properties": {
"active": {
"type": "boolean"
},
"capture_scope": {
"enum": [
"auto",
"window",
"desktop"
],
"type": "string"
},
"desktop_unlocked": {
"type": "boolean"
},
"effective_scope": {
"enum": [
"window",
"desktop"
],
"type": "string"
},
"escalation_detail": {
"anyOf": [
{
"type": "string"
},
{
"type": "null"
}
]
},
"escalation_reason": {
"anyOf": [
{
"enum": [
"ax_tree_pixel_mismatch",
"background_delivery_failed",
"foreground_ineffective",
"no_window_target",
"other"
],
"type": "string"
},
{
"type": "null"
}
]
},
"revived": {
"type": "boolean"
},
"session": {
"type": "string"
}
},
"required": [
"session",
"capture_scope",
"effective_scope",
"desktop_unlocked",
"escalation_reason",
"escalation_detail",
"active",
"revived"
],
"type": "object"
}
},
{
"name": "type_text",
"description": "Type text into the current foreground desktop application.",
"platforms": [
"macos",
"windows",
"linux"
],
"capabilities": [
"input.keyboard.type",
"input.keyboard.type.terminal_safe",
"accessibility.element_tokens"
],
"annotations": {
"read_only": false,
"destructive": true,
"idempotent": false,
"open_world": true
},
"schema_mode": "portable_subset",
"cursor_semantics": {
"action": "text"
},
"input_schema": {
"additionalProperties": false,
"properties": {
"scope": {
"description": "Deprecated flat desktop target retained for wire compatibility.",
"enum": [
"desktop"
],
"type": "string"
},
"session": {
"description": "For multi-call work, prefer a short public session label and repeat it on every call that\naccepts it. Omit it to use the authenticated transport's implicit lifecycle session.",
"type": "string"
},
"target": {
"additionalProperties": true,
"description": "Exact window or desktop target. Window requires pid and window_id; desktop requires display_id.",
"properties": {
"display_id": {
"description": "Use primary for the primary display.",
"type": "string"
},
"kind": {
"enum": [
"window",
"desktop"
],
"type": "string"
},
"pid": {
"minimum": 0,
"type": "integer"
},
"window_id": {
"minimum": 0,
"type": "integer"
}
},
"required": [
"kind"
],
"type": "object"
},
"text": {
"type": "string"
}
},
"required": [
"text"
],
"type": "object"
},
"success_output_schema": {
"additionalProperties": false,
"properties": {
"delivery": {
"additionalProperties": false,
"properties": {
"delivered_count": {
"minimum": 0,
"type": [
"integer",
"null"
]
},
"mode": {
"enum": [
"background",
"foreground",
"not_applicable",
"unknown"
],
"type": "string"
}
},
"required": [
"mode"
],
"type": [
"object",
"null"
]
},
"effect": {
"enum": [
"confirmed",
"partial",
"unverifiable",
"suspected_noop",
"refused"
],
"type": "string"
},
"escalation": {
"additionalProperties": false,
"properties": {
"reason": {
"enum": [
"route_unavailable",
"delivery_failed",
"effect_unconfirmed",
"suspected_noop",
"permission_required"
],
"type": "string"
},
"target": {
"enum": [
"pixel",
"foreground",
"page",
"session"
],
"type": "string"
}
},
"required": [
"target",
"reason"
],
"type": [
"object",
"null"
]
},
"evidence": {
"items": {
"additionalProperties": false,
"properties": {
"kind": {
"enum": [
"value_readback",
"window_change"
],
"type": "string"
}
},
"required": [
"kind"
],
"type": "object"
},
"type": [
"array",
"null"
]
},
"route": {
"enum": [
"accessibility",
"synthetic_events",
"global_input",
"system_api",
"dom",
"trusted_input"
],
"type": "string"
}
},
"required": [
"effect",
"route"
],
"type": "object"
}
},
{
"name": "verify_state",
"description": "Deterministically verify bounded predicates against one exact window. The driver evaluates structured window/accessibility state and may return the final screenshot as uninterpreted visual evidence for a multimodal caller. Predicate results are satisfied, unsatisfied, or unknown; unknown never implies success. Accessibility projections are conservative: absence remains unknown unless the observed search domain is proven exhaustive.",
"platforms": [
"macos",
"windows",
"linux"
],
"capabilities": [
"state.verify",
"state.verify.window",
"state.verify.element",
"state.observe.visual_evidence"
],
"annotations": {
"read_only": true,
"destructive": false,
"idempotent": false,
"open_world": false
},
"schema_mode": "portable_subset",
"input_schema": {
"additionalProperties": false,
"properties": {
"expect": {
"description": "One to eight predicates, combined with logical AND.",
"items": {
"additionalProperties": false,
"properties": {
"element": {
"additionalProperties": false,
"properties": {
"enabled": {
"type": "boolean"
},
"exists": {
"description": "Assert that at least one trusted element matches the selector.\n\nElement walks are not yet exhaustive on every platform, so absence\ncannot be proven. `false` is rejected instead of returning an\nindefinitely-unknown predicate.",
"type": "boolean"
},
"selected": {
"type": "boolean"
},
"selector": {
"additionalProperties": false,
"properties": {
"label_contains": {
"minLength": 1,
"type": "string"
},
"role": {
"minLength": 1,
"type": "string"
}
},
"required": [],
"type": "object"
},
"value_equals": {
"type": "string"
}
},
"required": [
"selector"
],
"type": "object"
},
"window": {
"additionalProperties": false,
"properties": {
"bounds": {
"additionalProperties": false,
"properties": {
"height": {
"type": "number"
},
"tolerance_px": {
"maximum": 100,
"minimum": 0,
"type": "number"
},
"width": {
"type": "number"
},
"x": {
"type": "number"
},
"y": {
"type": "number"
}
},
"required": [
"x",
"y",
"width",
"height"
],
"type": "object"
},
"exists": {
"type": "boolean"
}
},
"required": [],
"type": "object"
}
},
"required": [],
"type": "object"
},
"maxItems": 8,
"minItems": 1,
"type": "array"
},
"include_screenshot": {
"description": "Return the final window screenshot as image content for a multimodal\ncaller. The driver does not interpret that image.",
"type": "boolean"
},
"pid": {
"description": "Exact process whose window may be observed.",
"minimum": 1,
"type": "integer"
},
"session": {
"description": "For multi-call work, prefer a short public session label and repeat it on every call that\naccepts it. Omit it to use the authenticated transport's implicit lifecycle session. This\nfield never selects capture modality or authorization.",
"type": "string"
},
"stable_samples": {
"default": 2,
"description": "Consecutive satisfied samples required before returning success.",
"maximum": 5,
"minimum": 1,
"type": "integer"
},
"timeout_ms": {
"default": 5000,
"description": "Bounded wait. Zero performs one sample.",
"maximum": 10000,
"minimum": 0,
"type": "integer"
},
"window_id": {
"description": "Exact native window identifier.",
"type": "integer"
}
},
"required": [
"pid",
"window_id",
"expect"
],
"type": "object"
},
"success_output_schema": {
"additionalProperties": true,
"properties": {
"elapsed_ms": {
"minimum": 0,
"type": "integer"
},
"predicates": {
"items": {
"properties": {
"index": {
"minimum": 0,
"type": "integer"
},
"observed_json": {
"anyOf": [
{
"type": "string"
},
{
"type": "null"
}
]
},
"status": {
"enum": [
"satisfied",
"unsatisfied",
"unknown"
],
"type": "string"
},
"unknown_reason": {
"anyOf": [
{
"enum": [
"invalid_predicate",
"unsupported_predicate",
"untrusted_source",
"multi_match",
"target_missing",
"observation_unavailable",
"stability_unproven"
],
"type": "string"
},
{
"type": "null"
}
]
}
},
"required": [
"index",
"status",
"unknown_reason",
"observed_json"
],
"type": "object"
},
"type": "array"
},
"samples": {
"minimum": 0,
"type": "integer"
},
"stable": {
"type": "boolean"
},
"status": {
"enum": [
"satisfied",
"unsatisfied",
"unknown"
],
"type": "string"
}
},
"required": [
"status",
"stable",
"elapsed_ms",
"samples",
"predicates"
],
"type": "object"
}
}
]
}