pencil
Pass
Audited by Gen Agent Trust Hub on Oct 7, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill defines an ingestion surface for external data by reading and processing .pen design files.\n
- Ingestion points: Tools like open_document and batch_get in SKILL.md are used to read external file content into the agent's context.\n
- Boundary markers: The skill instructs the agent to treat these files as encrypted and strictly use tool-based access, which serves as a logical boundary for the agent.\n
- Capability inventory: The skill includes batch_design which allows file modification, and snapshot_layout/get_screenshot for analytical and visual output.\n
- Sanitization: The skill relies on the MCP server's internal handling of the design files rather than explicit prompt sanitization instructions.\n- [COMMAND_EXECUTION]: The MCP configuration reference provided in SKILL.md includes an absolute path to a local executable binary (/Users/wandl/.../mcp-server-darwin-arm64). This is a standard requirement for the stdio transport layer in MCP skills, although the absolute path makes it specific to the author's local environment.
Audit Metadata