The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill ingests data from external GitHub issues and local tracker documentation, which could contain instructions intended to influence agent behavior.\n- Ingestion points: Reads from GitHub issue titles, bodies, and comments, as well as docs/agents/issue-tracker.md.\n- Boundary markers: Relies on markdown headers and structured sections for organization, but lacks explicit delimiters to isolate untrusted content.\n- Capability inventory: Includes the ability to create and modify GitHub issues (labels, child tickets, assignments, blocking links, resolution comments) and create temporary research branches.\n- Sanitization: Ingested content is not specifically sanitized or filtered before being used to update the project map.

wayfinder