nestjs-feature-audit
Pass
Audited by Gen Agent Trust Hub on Oct 5, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill correctly separates auditing from implementation, ensuring that feature gaps are identified but not fixed without a separate, authorized phase. This prevents unauthorized repository mutations during the audit.
- [COMMAND_EXECUTION]: The skill executes Git and ripgrep commands for repository inspection. These are implemented safely with instructions to escape user-provided feature names or use fixed-string searches, mitigating command injection risks. Git usage is limited to non-destructive operations like
status,fetch, andmerge --ff-only, requiring user consent for non-standard worktree states. - [INDIRECT_PROMPT_INJECTION]: The skill processes external roadmap documentation, creating an ingestion surface for indirect prompt injection. Ingestion points: Documentation is read from Markdown, MDX, YAML, and JSON files in the
docs/directory or supplied directly by the user as text. Boundary markers: The skill uses a 'roadmap gate' to reject ambiguous plans and explicitly forbids treating roadmap text as permission to act. Capability inventory: The agent can run repository searches viaripgrepand safe Git operations. Sanitization: Instructions mandate escaping untrusted search terms or using fixed-string searches to prevent injection during processing.
Audit Metadata