submit-github-review
Pass
Audited by Gen Agent Trust Hub on Mar 11, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests an array of findings from specialist skills and interpolates the content directly into GitHub review summaries and inline comments. This constitutes a potential surface where malicious instructions within the findings could influence the agent or mislead reviewers.\n
- Ingestion points: The
findingsinput array defined inSKILL.md.\n - Boundary markers: Absent; the skill uses markdown templates but lacks specific delimiters or instructions to ignore embedded commands within the findings.\n
- Capability inventory: The skill utilizes the
create_pull_request_reviewMCP tool to post content to external GitHub repositories.\n - Sanitization: No explicit sanitization, escaping, or validation of the
findingscontent is defined in the skill logic.
Audit Metadata