{"passport":{"unfragile":{"@version":"1.0","version":"2026-05","artifact":{"id":"awesome-peekaboo","slug":"peekaboo","name":"Peekaboo","type":"mcp","url":"https://github.com/steipete/Peekaboo","page_url":"https://unfragile.ai/peekaboo","categories":["mcp-servers"],"tags":[],"pricing":{"model":"open_source","free":true,"starting_price":null},"status":"active","verified":false},"capabilities":[{"id":"awesome-peekaboo__cap_0","uri":"capability://image.visual.pixel.accurate.screen.capture.with.multi.display.and.window.scoped.targeting","name":"pixel-accurate screen capture with multi-display and window-scoped targeting","description":"Captures screenshots using ScreenCaptureKit (macOS 12.3+) with automatic CGWindow fallback, supporting Retina scaling (2x on HiDPI displays), multi-display targeting via screen index, window-scoped capture by app name/PID/window ID, and menu bar capture including status bar extras. The capture engine is abstraction-layered to allow runtime selection between ScreenCaptureKit and legacy CGWindow APIs based on availability and performance characteristics.","intents":["I need to capture the current state of a specific application window for AI vision analysis","I want to screenshot the entire system or specific displays for automated testing","I need to capture menu bar and status bar UI elements that are normally transient","I need high-fidelity pixel-perfect captures on Retina displays for visual inspection"],"best_for":["AI agents performing visual GUI automation on macOS","CI/CD pipelines running automated UI tests on macOS runners","Developers building macOS accessibility tools that need visual fallback when accessibility APIs fail"],"limitations":["macOS 12.3+ required for ScreenCaptureKit; older versions fall back to CGWindow which has lower fidelity","Requires Screen Recording permission in System Preferences; will fail silently if not granted","Menu bar capture is complex and may miss dynamically-rendered elements in third-party menu bar apps","Capture latency varies by display resolution and number of displays; HiDPI 2x scaling doubles memory usage"],"requires":["macOS 12.3+ (ScreenCaptureKit) or macOS 10.13+ (CGWindow fallback)","Screen Recording permission granted in System Preferences > Security & Privacy","Swift runtime (built into macOS)"],"input_types":["window identifier (app name, PID, or window ID)","screen index (integer for multi-display systems)","capture region (optional bounding box)"],"output_types":["PNG image data (in-memory or file)","JPEG image data (compressed)","raw pixel buffer (for downstream vision processing)"],"categories":["image-visual","automation-workflow"],"confidence":0.5,"matches":0,"success_rate":0},{"id":"awesome-peekaboo__cap_1","uri":"capability://data.processing.analysis.semantic.ui.element.detection.and.accessibility.based.interaction","name":"semantic ui element detection and accessibility-based interaction","description":"Detects interactive UI elements (buttons, text fields, menus, etc.) using macOS Accessibility APIs (AXUIElement) with fallback to vision-based element detection when accessibility metadata is unavailable. The system maintains a semantic element registry that maps detected elements to their accessibility attributes (role, label, value, enabled state) and enables deterministic interaction via native accessibility actions (click, type, select) rather than pixel-based mouse movement.","intents":["I want an AI agent to reliably click buttons and interact with UI elements without fragile coordinate-based automation","I need to detect what UI elements are available on screen and their current state (enabled/disabled, selected/unselected)","I want fallback vision-based element detection when an app doesn't expose accessibility metadata","I need to interact with dynamic UI that changes position or size between runs"],"best_for":["AI agents automating macOS applications with accessibility support (most modern apps)","Teams building robust RPA workflows that need to survive UI layout changes","Developers testing applications that have accessibility compliance requirements"],"limitations":["Accessibility APIs require Accessibility permission; apps without accessibility support fall back to vision-based detection which is slower and less reliable","Vision-based element detection adds ~200-500ms latency per detection pass and requires Claude vision or similar model","Some native macOS apps (e.g., certain menu bar apps, legacy Carbon apps) have incomplete accessibility metadata","Element detection snapshots are stored on disk and require cleanup; unbounded snapshot accumulation can consume significant storage"],"requires":["macOS 10.13+ (Accessibility APIs)","Accessibility permission granted for the Peekaboo process in System Preferences > Security & Privacy > Accessibility","Target application must expose AXUIElement hierarchy (most modern apps do)"],"input_types":["screen capture (PNG/JPEG from capture service)","accessibility element query (role, label, or attribute filter)","vision model context (for vision-based fallback detection)"],"output_types":["structured element list with attributes (role, label, value, position, enabled state)","element snapshots (visual crops of detected elements)","accessibility action registry (available actions per element)"],"categories":["data-processing-analysis","automation-workflow"],"confidence":0.5,"matches":0,"success_rate":0},{"id":"awesome-peekaboo__cap_10","uri":"capability://data.processing.analysis.snapshot.storage.and.cleanup.system.for.element.detection.artifacts","name":"snapshot storage and cleanup system for element detection artifacts","description":"Manages storage of element detection snapshots (visual crops of detected UI elements) on disk with automatic cleanup to prevent unbounded storage growth. The system stores snapshots in a configurable directory, tracks snapshot metadata (timestamp, element ID, size), and implements cleanup policies (age-based, size-based, LRU). Snapshots are used by vision models to analyze specific UI elements without re-capturing the entire screen.","intents":["I want to store visual crops of detected UI elements for vision model analysis","I need to prevent snapshot storage from consuming unlimited disk space","I want to debug element detection by inspecting stored snapshots","I need to manage snapshot lifecycle (creation, retention, deletion)"],"best_for":["Systems running long-lived Peekaboo agents that accumulate many snapshots","Teams with limited disk space who need automatic cleanup","Developers debugging element detection issues by inspecting snapshots"],"limitations":["Cleanup policies are not configurable at runtime; they require configuration file changes","Snapshot deletion is not reversible; deleted snapshots cannot be recovered","Cleanup runs synchronously; large cleanup operations may block agent execution","No snapshot compression; large snapshots consume significant disk space"],"requires":["Writable directory for snapshot storage (default: ~/.peekaboo/snapshots)","Disk space for snapshots (varies by usage; typical: 100MB-1GB per day)"],"input_types":["element detection results (with snapshot paths)","cleanup policy configuration"],"output_types":["snapshot file paths","cleanup statistics (deleted count, freed space)"],"categories":["data-processing-analysis","automation-workflow"],"confidence":0.5,"matches":0,"success_rate":0},{"id":"awesome-peekaboo__cap_11","uri":"capability://automation.workflow.mac.application.with.status.bar.interface.and.visual.inspector","name":"mac application with status bar interface and visual inspector","description":"Provides a native macOS application (Peekaboo.app) that runs in the status bar and offers a visual inspector for debugging Peekaboo operations. The app displays real-time screenshots, detected UI elements, and execution logs; allows users to manually trigger captures and interactions; and provides a settings interface for configuration. The app maintains a persistent connection to the Peekaboo service and streams events in real-time.","intents":["I want to visually inspect what Peekaboo is seeing and doing in real-time","I need a GUI to configure Peekaboo settings without editing configuration files","I want to manually test Peekaboo captures and interactions before automating them","I need to debug agent behavior by watching screenshots and element detection results"],"best_for":["Developers debugging Peekaboo automation interactively","Non-technical users who prefer GUI configuration over command-line setup","Teams building Peekaboo integrations who need a visual debugging tool"],"limitations":["Mac app is optional; core Peekaboo functionality works without it","Real-time event streaming adds overhead; running the app may slightly slow down automation","Settings changes in the app may not be reflected in CLI or MCP server until restart","Visual inspector is read-only; it cannot directly trigger complex automation sequences"],"requires":["macOS 12.3+ (for full feature support)","Peekaboo service running in background","Screen Recording permission (for screenshot display)"],"input_types":["manual trigger actions (capture, detect elements, interact)","settings changes (via GUI)"],"output_types":["real-time screenshots","detected UI elements with metadata","execution logs and events","settings configuration"],"categories":["automation-workflow"],"confidence":0.5,"matches":0,"success_rate":0},{"id":"awesome-peekaboo__cap_12","uri":"capability://automation.workflow.speech.recognition.integration.for.voice.based.interaction","name":"speech recognition integration for voice-based interaction","description":"Integrates macOS native speech recognition (via Speech framework) to enable voice-based interaction with the Peekaboo agent. The system captures audio input, transcribes it to text using on-device speech recognition, and passes the transcribed text to the agent as a natural language instruction. Speech recognition runs asynchronously and supports real-time transcription feedback.","intents":["I want to control macOS automation using voice commands instead of typing","I need hands-free interaction with the Peekaboo agent for accessibility","I want to provide natural language instructions via microphone input"],"best_for":["Accessibility-focused users who need hands-free automation","Users in environments where typing is inconvenient (e.g., while using other applications)","Teams building voice-controlled macOS assistants"],"limitations":["Speech recognition accuracy depends on audio quality and background noise; noisy environments may produce incorrect transcriptions","On-device speech recognition has limited language support; some languages may not be available","Microphone permission is required; users must grant permission in System Preferences","Real-time transcription feedback adds latency; there is a delay between speech and text output"],"requires":["macOS 10.15+ (Speech framework)","Microphone permission granted in System Preferences > Security & Privacy > Microphone","Microphone hardware"],"input_types":["audio input from microphone"],"output_types":["transcribed text","transcription confidence score","real-time transcription updates"],"categories":["automation-workflow"],"confidence":0.5,"matches":0,"success_rate":0},{"id":"awesome-peekaboo__cap_13","uri":"capability://automation.workflow.error.handling.system.with.diagnostic.reporting.and.recovery.strategies","name":"error handling system with diagnostic reporting and recovery strategies","description":"Implements a comprehensive error handling system that captures detailed diagnostic information (stack traces, system state, screenshots) when operations fail, provides human-readable error messages, and implements recovery strategies (retry with backoff, fallback paths, state rollback). The system categorizes errors by severity and type, enabling targeted recovery logic and diagnostic reporting.","intents":["I want detailed diagnostic information when Peekaboo operations fail","I need automatic retry logic for transient failures without manual intervention","I want to understand why an automation failed and what went wrong","I need fallback strategies when primary automation paths fail"],"best_for":["Teams running Peekaboo in production who need reliable error recovery","Developers debugging complex automation failures","Organizations that need detailed error logs for compliance or troubleshooting"],"limitations":["Error recovery is limited to predefined strategies; novel error scenarios may not have recovery logic","Diagnostic information (screenshots, logs) can consume significant disk space","Retry logic uses exponential backoff which may be too slow for time-sensitive operations","Some errors (e.g., permission denied) cannot be recovered automatically"],"requires":["Writable directory for error logs and diagnostics (default: ~/.peekaboo/logs)","Disk space for diagnostic artifacts (varies by error frequency)"],"input_types":["operation that failed","error context (system state, screenshots, logs)"],"output_types":["error message (human-readable)","error code (for programmatic handling)","diagnostic report (stack trace, system state, screenshots)","recovery action (retry, fallback, abort)"],"categories":["automation-workflow","safety-moderation"],"confidence":0.5,"matches":0,"success_rate":0},{"id":"awesome-peekaboo__cap_2","uri":"capability://automation.workflow.deterministic.ui.interaction.via.accessibility.actions.and.synthetic.input","name":"deterministic ui interaction via accessibility actions and synthetic input","description":"Executes deterministic UI interactions (click, type, select, scroll, drag) using native macOS accessibility actions (AXPress, AXSetValue, etc.) when elements expose accessibility metadata, with fallback to synthetic input (CGEvent-based mouse/keyboard events) for inaccessible elements. The system maintains an interaction queue that serializes actions to prevent race conditions and includes error recovery logic that retries failed interactions with exponential backoff.","intents":["I want an AI agent to click a button, type text, or select a menu item reliably without coordinate fragility","I need to handle interaction failures gracefully with automatic retry and error reporting","I want to perform complex multi-step interactions (e.g., fill form, submit, wait for result) atomically","I need to interact with elements that move or resize dynamically between detection and interaction"],"best_for":["AI agents automating macOS applications in production workflows","CI/CD pipelines running automated UI tests that need to survive minor layout changes","Teams building RPA solutions that require high reliability and error recovery"],"limitations":["Accessibility actions require the target application to implement AXPress/AXSetValue; some legacy apps only support synthetic input which is slower and more fragile","Synthetic input (CGEvent) requires Universal Access permission and may be blocked by security features in newer macOS versions","Interaction latency varies by application; some apps have slow event processing that causes timeouts","No built-in transaction support; multi-step interactions can partially fail, leaving the UI in an inconsistent state"],"requires":["macOS 10.13+ (Accessibility APIs) or macOS 10.12+ (CGEvent synthetic input)","Accessibility permission for Peekaboo process","Target application must be in foreground or support background interaction (varies by app)"],"input_types":["interaction command (click, type, select, scroll, drag with parameters)","target element (identified by accessibility element reference or coordinates)","input data (text to type, scroll distance, drag coordinates)"],"output_types":["interaction result (success/failure with error code)","post-interaction state (screenshot or element snapshot)","interaction latency metrics (for performance monitoring)"],"categories":["automation-workflow","tool-use-integration"],"confidence":0.5,"matches":0,"success_rate":0},{"id":"awesome-peekaboo__cap_3","uri":"capability://automation.workflow.window.and.space.management.with.heuristic.based.window.selection","name":"window and space management with heuristic-based window selection","description":"Manages macOS window lifecycle and space (virtual desktop) navigation using a heuristic-based window selection system that ranks windows by relevance (foreground status, recent focus, window type). The system can enumerate all windows, filter by application, activate windows, move windows between spaces, and handle window-scoped operations. Window selection heuristics account for hidden windows, minimized windows, and multiple windows from the same application.","intents":["I want to activate a specific application window or switch between windows programmatically","I need to move a window to a different space (virtual desktop) or arrange windows on screen","I want to enumerate all open windows and filter by application or window properties","I need to handle cases where an application has multiple windows and select the most relevant one"],"best_for":["AI agents that need to manage window focus and visibility for multi-application workflows","Automation scripts that coordinate work across multiple spaces or virtual desktops","Teams building macOS workspace management tools"],"limitations":["Window selection heuristics are probabilistic; in edge cases with many similar windows, the wrong window may be selected","Space (virtual desktop) operations require Mission Control to be enabled; some users disable it","Some applications (e.g., certain menu bar apps, full-screen apps) don't expose standard window APIs","Window focus changes are asynchronous; there is no guarantee that a window will be in foreground immediately after activation"],"requires":["macOS 10.13+ (window management APIs)","Accessibility permission for Peekaboo process","Mission Control enabled (for space navigation)"],"input_types":["application name or bundle identifier","window ID or window title","space index (for space navigation)","window filter criteria (visible, minimized, etc.)"],"output_types":["window list with metadata (ID, title, app, position, size, space)","selected window reference","operation result (success/failure)"],"categories":["automation-workflow"],"confidence":0.5,"matches":0,"success_rate":0},{"id":"awesome-peekaboo__cap_4","uri":"capability://automation.workflow.application.lifecycle.control.and.menu.dock.automation","name":"application lifecycle control and menu/dock automation","description":"Controls application lifecycle (launch, quit, force quit) and automates menu bar and dock interactions using a combination of accessibility APIs and synthetic input. The system can launch applications by bundle identifier or path, terminate applications gracefully or forcefully, interact with menu items (click, check state), and manipulate dock icons. Menu automation includes special handling for hierarchical menus and dynamic menu items that appear/disappear based on application state.","intents":["I want to launch an application and wait for it to be ready for interaction","I need to click menu items or check menu state programmatically","I want to interact with dock icons (open, close, show/hide)","I need to force-quit an unresponsive application or gracefully terminate it"],"best_for":["AI agents orchestrating multi-application workflows that require app lifecycle management","Automation scripts that need to interact with menu-driven applications","CI/CD pipelines that need to launch and tear down macOS applications"],"limitations":["Menu automation is fragile because menu items are dynamically rendered and may not expose accessibility metadata","Dock automation is limited; some dock features (e.g., drag-and-drop) are not reliably automatable","Application launch is asynchronous; there is no reliable way to detect when an app is fully ready for interaction","Force quit may cause data loss; graceful quit may hang if the application is unresponsive"],"requires":["macOS 10.13+ (application control APIs)","Accessibility permission for Peekaboo process","Target application must be installed and accessible"],"input_types":["application bundle identifier or path","menu item path (e.g., 'File > Open')","dock icon identifier","quit mode (graceful or force)"],"output_types":["application process ID (PID)","menu item state (checked, enabled, etc.)","operation result (success/failure)"],"categories":["automation-workflow"],"confidence":0.5,"matches":0,"success_rate":0},{"id":"awesome-peekaboo__cap_5","uri":"capability://tool.use.integration.mcp.server.integration.for.ai.client.connectivity","name":"mcp server integration for ai client connectivity","description":"Exposes Peekaboo capabilities as an MCP (Model Context Protocol) server that can be connected to AI clients (Claude Desktop, Cursor, other MCP-compatible tools) via stdio or HTTP transport. The MCP server implements a tool registry that maps Peekaboo commands to MCP tool definitions with JSON schema validation, handles async tool execution with streaming output, and manages session state across multiple tool calls. The server supports both synchronous and asynchronous tool execution patterns.","intents":["I want to connect Claude Desktop or Cursor to Peekaboo so the AI can automate my macOS applications","I need to expose Peekaboo as a service that multiple AI clients can connect to simultaneously","I want to build a custom AI agent that uses Peekaboo tools via MCP protocol","I need to integrate Peekaboo into an existing MCP-based AI orchestration system"],"best_for":["AI application developers building Claude Desktop or Cursor plugins","Teams deploying Peekaboo as a shared service for multiple AI agents","Developers building custom AI agents that need macOS GUI automation"],"limitations":["MCP server requires Node.js runtime; the underlying Peekaboo service is Swift-based, requiring a bridge process","Stdio transport has limited bandwidth; large screenshot transfers may be slow","HTTP transport requires additional security configuration (authentication, TLS)","Session state is not persisted; restarting the MCP server loses context"],"requires":["Node.js 18+ (for MCP server runtime)","Peekaboo CLI or Mac App running as background service","MCP-compatible AI client (Claude Desktop, Cursor, etc.)"],"input_types":["MCP tool call with JSON parameters","streaming input (for long-running operations)"],"output_types":["MCP tool result with JSON response","streaming output (screenshots, logs, progress updates)","error responses with diagnostic information"],"categories":["tool-use-integration","automation-workflow"],"confidence":0.5,"matches":0,"success_rate":0},{"id":"awesome-peekaboo__cap_6","uri":"capability://planning.reasoning.ai.agent.execution.pipeline.with.tool.system.and.model.selection","name":"ai agent execution pipeline with tool system and model selection","description":"Implements a complete AI agent execution pipeline that orchestrates tool calls, manages session state, and supports multiple LLM models via a model selection system (Tachikoma). The pipeline includes tool registration, JSON schema validation, execution queuing, error handling with retry logic, and event streaming for real-time progress updates. The system supports both interactive chat mode (multi-turn conversation) and batch execution mode (single-shot automation).","intents":["I want to run an AI agent that can autonomously complete macOS automation tasks using Peekaboo tools","I need to support multiple LLM models (OpenAI, Anthropic, local models) without changing agent code","I want to monitor agent execution in real-time with streaming output and progress updates","I need to handle agent failures gracefully with retry logic and error recovery"],"best_for":["Teams building autonomous AI agents for macOS automation","Developers who want to switch between LLM providers without code changes","Organizations running Peekaboo as a service with multiple concurrent agents"],"limitations":["Agent execution is stateless across restarts; no built-in persistence of agent state or conversation history","Tool execution is serialized; parallel tool calls are not supported, limiting throughput","Model selection (Tachikoma) adds complexity; configuration errors can cause agent failures","Error recovery is limited to retry logic; complex error scenarios may require manual intervention"],"requires":["LLM API key (OpenAI, Anthropic, or local model endpoint)","Configuration file with model selection and credentials","Peekaboo CLI or MCP server running"],"input_types":["task description (natural language)","tool registry (available Peekaboo tools)","model configuration (provider, model name, parameters)"],"output_types":["agent execution trace (tool calls, results, reasoning)","final result (success/failure with output)","event stream (real-time progress updates)"],"categories":["planning-reasoning","automation-workflow","tool-use-integration"],"confidence":0.5,"matches":0,"success_rate":0},{"id":"awesome-peekaboo__cap_7","uri":"capability://automation.workflow.configuration.system.with.credentials.management.and.permissions.tracking","name":"configuration system with credentials management and permissions tracking","description":"Provides a hierarchical configuration system that manages Peekaboo settings, LLM credentials, and macOS permissions state. The system supports configuration files (YAML/JSON), environment variables, and credential storage via secure macOS Keychain integration. The permissions tracking system monitors which macOS permissions have been granted (Screen Recording, Accessibility, etc.) and provides diagnostic information when permissions are missing.","intents":["I want to configure Peekaboo with LLM credentials and model preferences without hardcoding secrets","I need to check which macOS permissions are required and whether they have been granted","I want to use environment variables to override configuration for CI/CD pipelines","I need to store sensitive credentials securely using macOS Keychain"],"best_for":["Teams deploying Peekaboo in CI/CD pipelines with environment-based configuration","Developers building Peekaboo integrations who need secure credential storage","Organizations managing multiple Peekaboo instances with different configurations"],"limitations":["Configuration file format is not validated at load time; invalid YAML/JSON will cause runtime errors","Keychain integration is macOS-specific; no cross-platform credential storage","Permissions tracking is read-only; the system cannot grant permissions programmatically","Environment variable overrides are not documented; discovering available overrides requires code inspection"],"requires":["Configuration file in YAML or JSON format (optional; defaults can be used)","macOS Keychain for credential storage (automatic on macOS)","Environment variables for CI/CD overrides (optional)"],"input_types":["configuration file path","environment variable names","credential key names"],"output_types":["parsed configuration object","permissions status report","diagnostic messages for missing permissions"],"categories":["automation-workflow"],"confidence":0.5,"matches":0,"success_rate":0},{"id":"awesome-peekaboo__cap_8","uri":"capability://automation.workflow.cli.command.interface.with.vision.and.interaction.subcommands","name":"cli command interface with vision and interaction subcommands","description":"Provides a comprehensive CLI interface (peekaboo binary) that exposes all Peekaboo capabilities as subcommands organized by category: vision commands (screenshot, detect-elements), interaction commands (click, type, select), window management commands (list-windows, activate-window), application control commands (launch, quit), menu automation commands (click-menu), and system utility commands. Each command supports JSON output for programmatic parsing and includes detailed help text.","intents":["I want to script macOS automation using shell commands in CI/CD pipelines","I need to test Peekaboo capabilities interactively from the command line","I want to build custom automation scripts that combine Peekaboo commands with shell logic","I need to integrate Peekaboo into existing automation frameworks that expect CLI tools"],"best_for":["DevOps engineers building CI/CD pipelines with macOS test runners","Shell script developers who want to automate macOS GUIs","Teams integrating Peekaboo into existing automation frameworks"],"limitations":["CLI output is line-based; large screenshots or complex data structures are difficult to parse","JSON output mode is available but not all commands support it; some commands only support human-readable output","Error handling is basic; exit codes don't distinguish between different error types","No built-in command composition; complex workflows require shell scripting"],"requires":["Peekaboo CLI binary installed and in PATH","Shell environment (bash, zsh, etc.)","macOS 12.3+ (for full feature support)"],"input_types":["command name and arguments","flags (--json, --verbose, etc.)","stdin (for piping data between commands)"],"output_types":["human-readable text output","JSON output (for programmatic parsing)","exit codes (0 for success, non-zero for failure)"],"categories":["automation-workflow","tool-use-integration"],"confidence":0.5,"matches":0,"success_rate":0},{"id":"awesome-peekaboo__cap_9","uri":"capability://planning.reasoning.interactive.chat.mode.with.multi.turn.conversation.and.session.management","name":"interactive chat mode with multi-turn conversation and session management","description":"Implements an interactive chat interface that enables multi-turn conversations with an AI agent, maintaining session state across multiple user inputs. The system manages conversation history, tool execution context, and screenshot state, allowing users to give natural language instructions and receive feedback in real-time. The chat mode supports both CLI-based interaction and programmatic session management via the Agent API.","intents":["I want to have a natural language conversation with an AI agent to automate macOS tasks","I need to provide feedback or corrections to the agent mid-execution without restarting","I want to maintain context across multiple automation steps without re-explaining the task","I need to inspect agent reasoning and screenshots during execution for debugging"],"best_for":["Interactive users who prefer conversational interfaces over scripting","Developers debugging agent behavior by inspecting intermediate states","Teams building AI-powered macOS assistants with conversational UX"],"limitations":["Session state is not persisted; closing the chat loses conversation history","Multi-turn conversations accumulate token usage; long conversations become expensive with paid LLM APIs","User feedback is not automatically incorporated; the agent may repeat mistakes if not explicitly corrected","Chat mode is single-user; concurrent chat sessions are not supported"],"requires":["LLM API key (OpenAI, Anthropic, or local model endpoint)","Interactive terminal or chat interface","Peekaboo service running in background"],"input_types":["natural language user input","feedback on agent actions (approve/reject/modify)"],"output_types":["agent reasoning and planning","tool execution results","screenshots and element detection results","final task completion status"],"categories":["planning-reasoning","automation-workflow"],"confidence":0.5,"matches":0,"success_rate":0}],"trust":{"score":32,"verified":false,"data_access_risk":"high","permissions":["macOS 12.3+ (ScreenCaptureKit) or macOS 10.13+ (CGWindow fallback)","Screen Recording permission granted in System Preferences > Security & Privacy","Swift runtime (built into macOS)","macOS 10.13+ (Accessibility APIs)","Accessibility permission granted for the Peekaboo process in System Preferences > Security & Privacy > Accessibility","Target application must expose AXUIElement hierarchy (most modern apps do)","Writable directory for snapshot storage (default: ~/.peekaboo/snapshots)","Disk space for snapshots (varies by usage; typical: 100MB-1GB per day)","macOS 12.3+ (for full feature support)","Peekaboo service running in background"],"failure_modes":["macOS 12.3+ required for ScreenCaptureKit; older versions fall back to CGWindow which has lower fidelity","Requires Screen Recording permission in System Preferences; will fail silently if not granted","Menu bar capture is complex and may miss dynamically-rendered elements in third-party menu bar apps","Capture latency varies by display resolution and number of displays; HiDPI 2x scaling doubles memory usage","Accessibility APIs require Accessibility permission; apps without accessibility support fall back to vision-based detection which is slower and less reliable","Vision-based element detection adds ~200-500ms latency per detection pass and requires Claude vision or similar model","Some native macOS apps (e.g., certain menu bar apps, legacy Carbon apps) have incomplete accessibility metadata","Element detection snapshots are stored on disk and require cleanup; unbounded snapshot accumulation can consume significant storage","Cleanup policies are not configurable at runtime; they require configuration file changes","Snapshot deletion is not reversible; deleted snapshots cannot be recovered","builder identity is not verified yet","no observed match outcomes yet"],"rank_breakdown":{"adoption":0.05,"quality":0.5,"ecosystem":0.39999999999999997,"match_graph":0.25,"freshness":0.52,"weights":{"adoption":0.25,"quality":0.25,"ecosystem":0.15,"match_graph":0.23,"freshness":0.12}},"observed_outcomes":{"matches":0,"success_rate":0,"avg_confidence":0,"top_intents":[],"last_matched_at":null},"maintenance":{"status":"active","updated_at":"2026-06-17T09:51:03.579Z","last_scraped_at":"2026-05-03T14:00:15.503Z","last_commit":null},"community":{"stars":null,"forks":null,"weekly_downloads":null,"model_downloads":null,"model_likes":null}},"distribution":{"claim_url":"https://unfragile.ai/submit?claim=peekaboo","compare_url":"https://unfragile.ai/compare?artifact=peekaboo"}},"signature":"/na1acKx+CHoJIrkKL0kGhopMAts/mTY9+B+E1H1VpBj3hl0FYZNQqOITHE41ZQhCm/JrA6p7HpA2L9zyNP1DA==","signedAt":"2026-06-20T15:16:23.517Z","signedBy":"unfragile.ai","version":1},"_links":{"self":"https://unfragile.ai/api/v1/passport/peekaboo","artifact":"https://unfragile.ai/peekaboo","verify":"https://unfragile.ai/api/v1/verify?slug=peekaboo","publicKey":"https://unfragile.ai/api/v1/trust-passport-public-key","spec":"https://unfragile.ai/trust","schema":"https://unfragile.ai/schema.json","docs":"https://unfragile.ai/docs"}}