clean-comments
Pass
Audited by Gen Agent Trust Hub on Nov 4, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTION
Full Analysis
- [COMMAND_EXECUTION]: The skill executes
git statusto verify the repository state and runs a bundled Node.js script (scripts/build-report.mjs) to generate a final report.\n- [INDIRECT_PROMPT_INJECTION]: The skill reads and processes code comments from the user's repository, which serves as an ingestion point for untrusted data.\n - Ingestion points: The agent reads file content from the local repository as specified by the user.\n
- Boundary markers: The instructions do not define explicit markers to isolate analyzed code content, which could allow instructions in comments to influence agent behavior.\n
- Capability inventory: The agent has the ability to rewrite files in the repository and execute shell commands to run reporting scripts.\n
- Sanitization: The
build-report.mjsscript escapes metadata values and uses the establishedhighlight.jslibrary for rendering code context.\n- [DYNAMIC_EXECUTION]: The skill uses a bundled script to generate an HTML report, utilizing Node.js built-ins and a local version ofhighlight.jsfor syntax highlighting.
Audit Metadata