final-release-review
Pass
Audited by Gen Agent Trust Hub on Sep 22, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- Indirect Prompt Injection Surface: The skill processes untrusted text from git logs, commit messages, and pull request descriptions to generate release reports. This creates a potential surface where specially crafted commit messages could attempt to influence the agent's ship/block recommendation.\n
- Ingestion points: External data enters the context during the diff snapshot and documentation review steps in
SKILL.mdandreferences/review-checklist.md.\n - Boundary markers: The skill uses a specific Markdown template for its output report, which helps contain the influence of ingested text.\n
- Capability inventory: The skill performs local git read operations and utilizes read-only GitHub API access; it does not have capabilities for file modification or external data exfiltration.\n
- Sanitization: The skill relies on structured auditing steps and a deterministic gate policy rather than explicit string sanitization for ingested git data.
Audit Metadata