project-state-governor
Pass
Audited by Gen Agent Trust Hub on Sep 20, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill instructs the agent to ingest and analyze data from the repository (code, documentation, tests, and history) to update project state files. This process creates a vulnerability to indirect prompt injection where malicious repository content could mislead the agent.
- Ingestion points: The workflow defined in
SKILL.md(Section 5) andreferences/reconstruction-workflow.md(Section 2) requires the agent to read repository evidence, documentation, and Git history. - Boundary markers: The skill provides a defense-in-depth approach through an "Authority hierarchy" (Section 1) and a "Persistence lifecycle" (Section 8) that requires verification against objective Definitions of Done (DoD) before state is updated.
- Capability inventory: The agent is authorized to perform file and directory operations, including creating, modifying, and deleting documentation and state files (
SKILL.mdSection 14). - Sanitization: Section 12 explicitly directs the agent to avoid persisting secrets, credentials, or sensitive personal data during its state maintenance tasks.
Audit Metadata