plan
Pass
Audited by Gen Agent Trust Hub on Sep 23, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill determines project goals and scope by ingesting data from external sources including user requests, existing project records, issue trackers, and specifications. This establishes a surface for indirect prompt injection where malicious instructions hidden in these external sources could influence the generated plans or subsequent agent actions.\n
- Ingestion points: SKILL.md (Determines goals and scope using user requests, approved Specs, issues, and existing project records).\n
- Boundary markers: Absent (No specific delimiters or instructions to ignore embedded commands are specified for the external data).\n
- Capability inventory: SKILL.md (Performs file writes to
docs/plans/,.harness/state.md,.harness/work_index.md, andAGENTS.md; deletes artifact files during retirement; and creates git commits).\n - Sanitization: Absent (The workflow does not define validation or sanitization steps for data imported from external systems).
Audit Metadata