codereview-scan
Pass
Audited by Gen Agent Trust Hub on Oct 8, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill instructs the agent to audit entire repositories, directories, or specific files using the
ocr scancommand. This creates a surface for indirect prompt injection where instructions embedded in the source code being scanned (such as within comments or strings) could potentially influence the agent's analysis or reporting. - Ingestion points: Source code files read from the paths provided to the
ocr scan --path <path>command inSKILL.mdandreferences/scan-scenarios.md. - Boundary markers: No explicit instructions are provided to the agent to use delimiters or to treat the content of the scanned files as data rather than instructions.
- Capability inventory: The agent executes CLI commands (
ocr,open-code-review,codereview-harness) and generates human-readable reports based on the tool outputs. - Sanitization: No sanitization or filtering of the scanned file content is specified.
Audit Metadata