wayfinder
Pass
Audited by Gen Agent Trust Hub on Jul 12, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection (Category 8) due to its core functionality of reading and acting upon instructions within project map files.
- Ingestion points: The skill explicitly supports loading a 'map.md' file from a user-provided URL during the 'Work through the map' flow.
- Boundary markers: The instructions lack clear boundaries or warnings for the agent to treat the 'Notes' or 'Destination' sections of the map as untrusted data. The 'Notes' section specifically tells the agent to 'invoke the skills the map's Notes name'.
- Capability inventory: The skill has the ability to read and write files in the local
.ai/wayfinder/directory and can invoke other skills likegrilling,domain-modeling, anddeep-plan. - Sanitization: There is no evidence of sanitization or structure validation for the map content before it influences the agent's logic.
Audit Metadata