wayfinder
Pass
Audited by Gen Agent Trust Hub on Aug 10, 2026
Risk Level: SAFECOMMAND_EXECUTIONDATA_EXFILTRATIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill executes shell scripts that perform local file operations such as creating directories and writing project metadata within a dedicated directory. These actions are scoped to the project environment.\n- [DYNAMIC_EXECUTION]: Scripts use embedded Python to perform JSON operations and validate the ticket graph, which is a standard method for implementing logic in shell-based skills.\n- [INDIRECT_PROMPT_INJECTION]: The skill aggregates user-provided content from tickets into a final plan document, creating a potential path for instruction injection if the plan is consumed by other agents.\n
- Ingestion points: Markdown ticket files read by the wayfinder-close.sh script.\n
- Boundary markers: The generated plan uses headers but does not include explicit isolation markers for the aggregated content.\n
- Capability inventory: The skill has the ability to write files and append logs within the local project metadata directory.\n
- Sanitization: Content is extracted using regex and interpolated into documents without specific escaping or sanitization.
Audit Metadata