quality-gate
Pass
Audited by Gen Agent Trust Hub on Oct 9, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from external sources, specifically code diffs and GitHub Pull Request content, creating a surface for indirect prompt injection.
- Ingestion points: Untrusted data enters the skill's context through
git diff,gh pr diff, andgh pr viewcommands in Step 1 and Step 2. - Boundary markers: The prompt construction in Step 3 for the
majestic-engineer:workflow:quality-gatetask lacks explicit delimiters or "ignore previous instructions" markers to isolate the analyzed code changes from the agent's instructions. - Capability inventory: The skill has access to the
BashandTasktools, which are used to retrieve code and execute the review workflow. - Sanitization: No sanitization, escaping, or filtering logic is applied to the retrieved code or PR descriptions before they are interpolated into the prompt for the quality gate agent.
Audit Metadata