work-router
Pass
Audited by Gen Agent Trust Hub on Sep 30, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill handles raw user requests to determine the next skill in a project pipeline, creating a surface for indirect instructions.\n- Ingestion points: The user's request text is the primary input processed by the routing logic in
SKILL.md.\n- Boundary markers: Absent; there are no instructions to the agent to treat the user request as untrusted data or to use delimiters.\n- Capability inventory: The skill generates routing instructions that trigger other specialized skills with capabilities like GitHub issue creation and code modification.\n- Sanitization: No validation or escaping is performed on the user request content before intent resolution.
Audit Metadata