vet
Pass
Audited by Gen Agent Trust Hub on Aug 31, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill uses Git and GitHub CLI commands (such as
git rev-list,git cat-file,git diff, andgh run list) to audit the state of a repository. These commands are used for inspection purposes and do not modify the system or repository state. - [INDIRECT_PROMPT_INJECTION]: The skill ingests data from external sources, specifically Git logs and GitHub CI statuses.
- Ingestion points: Data enters the context through
git log,git rev-list, andgh run listoutput inSKILL.mdStep 1. - Boundary markers: None present; the skill treats command output as direct data for the report.
- Capability inventory: The skill uses Git and GitHub CLI tools for read-only operations. It does not perform file-writes, arbitrary code execution, or non-whitelisted network operations.
- Sanitization: No specific sanitization or filtering of commit messages or branch names is mentioned.
- Risk Assessment: The risk is minimal because the skill is explicitly read-only and its purpose is to generate a report for a human operator.
Audit Metadata