cloudbase-code-review
Pass
Audited by Gen Agent Trust Hub on Sep 8, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes external source code files which presents an indirect prompt injection surface where malicious instructions could be hidden in code comments or strings. However, the skill only performs analysis and reporting without executing high-privilege actions based on the ingested content.\n
- Ingestion points: The skill and its associated lint script read source files (.ts, .js, .html, .sql, .json) from the project directory.\n
- Boundary markers: Analysis is delimited by specific "LLM Review" sections and structured rule categories.\n
- Capability inventory: The skill uses file-read capabilities and a user-executed lint script that performs regex-based analysis.\n
- Sanitization: The process relies on deterministic regex patterns to identify issues, reducing the impact of injection.\n- [COMMAND_EXECUTION]: The skill provides a Node.js script in
references/lint-rules/README.mdfor manual execution by the user. The script is transparently provided and performs read-only file operations to validate project code and configuration against known security rules.
Audit Metadata