sketch
Pass
Audited by Gen Agent Trust Hub on Aug 18, 2026
Risk Level: SAFECOMMAND_EXECUTIONDATA_EXFILTRATIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The Preflight section of SKILL.md instructs the user or agent to run 'pkill -f sketch-mcp-server/dist/index.js' to manage stale server processes. This identifies a process termination capability within the skill's operational context.\n- [DATA_EXFILTRATION]: The tool 'sketch_capture_screenshot' returns a PNG of the current canvas. While functional, it allows for the visual exfiltration of any sensitive data that may be rendered or edited on the canvas.\n- [INDIRECT_PROMPT_INJECTION]: The skill processes external data through canvas reading and template loading.\n
- Ingestion points: Tools 'sketch_get_svg' and 'sketch_load_template' read data from the browser state and local JSON files respectively.\n
- Boundary markers: The instructions lack explicit boundary markers or warnings to ignore embedded instructions in the ingested SVG or JSON content.\n
- Capability inventory: The skill possesses capabilities for shell command execution (pkill), file system access (templates), and network synchronization (WebSockets).\n
- Sanitization: There is no mention of sanitization, validation, or filtering of the external content before it is processed by the agent.
Audit Metadata