wayfinder-with-visuals
Pass
Audited by Gen Agent Trust Hub on Nov 10, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted input from 'tickets' and 'maps' to determine workflow logic. There is an absence of explicit boundary markers or sanitization logic to prevent embedded instructions in this data from being interpreted by the agent.
- [INDIRECT_PROMPT_INJECTION]: The skill dynamically resolves and reads instruction sets (SKILL.md) from the local file system at runtime (e.g., ~/.claude/skills/ or ~/.agents/skills/). This creates a dependency on the integrity of local files to drive agent actions.
- [COMMAND_EXECUTION]: The instructions direct the agent to 'run' and 'follow' external skills based on the content of tickets, effectively delegating execution logic to the logic defined in those external dependencies.
Audit Metadata