elite-review
Pass
Audited by Gen Agent Trust Hub on Apr 13, 2026
Risk Level: SAFECOMMAND_EXECUTION
Full Analysis
- [COMMAND_EXECUTION]: The skill instructs the agent to execute local shell commands such as
git diffandgh pr diffto retrieve code changes for analysis. This is a standard and expected functionality for a code review tool. - [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data in the form of code diffs and PR content.
- Ingestion points: Data is ingested via
gh pr diff,git diff, and direct file reads as specified inSKILL.md. - Boundary markers: The instructions do not specify explicit boundary markers or delimiters to isolate the code being reviewed from the agent's instructions.
- Capability inventory: The agent's capabilities are limited to reading files, executing git/gh commands, and outputting a markdown report. It does not possess high-risk capabilities like arbitrary code execution or network exfiltration.
- Sanitization: There is no mention of sanitizing or escaping the content of the diffs before processing.
- Note: This represents a standard surface for indirect prompt injection where a malicious PR could contain comments or code specifically crafted to influence the AI's review verdict.
Audit Metadata