dev-review
Pass
Audited by Gen Agent Trust Hub on May 4, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill executes
ghCLI commands (pr list,pr view,pr diff) to interact with the repository's pull requests. These operations are essential for the skill's function and are limited to read-only access. - [PROMPT_INJECTION]: The skill has a surface for indirect prompt injection as it ingests and processes untrusted data from GitHub pull requests. An attacker could potentially include malicious instructions in a PR description or diff to manipulate the agent's analysis.
- Ingestion points: Content is ingested from
gh pr viewandgh pr diffinSKILL.md. - Boundary markers: The instructions do not define clear boundaries or 'ignore' directives for the content retrieved from external sources.
- Capability inventory: The skill is configured to use standard GitHub CLI tools for reading data.
- Sanitization: There is no evidence of filtering or sanitizing the content fetched from the pull requests before processing.
Audit Metadata