housekeep-gates
Pass
Audited by Gen Agent Trust Hub on Sep 13, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes CI configuration files and mapping data, creating a surface where malicious instructions within those files could influence agent actions during consolidation. \n
- Ingestion points: Repository CI workflows and archaeology maps referenced in Phase 0. \n
- Boundary markers: None explicitly defined. \n
- Capability inventory: Instructions to delete configuration files and modify branch protection settings in Phase 1 and 2. \n
- Sanitization: The skill uses post-execution probes (Phase 4) rather than input sanitization.
Audit Metadata