git-codebase-preflight
Pass
Audited by Gen Agent Trust Hub on Oct 3, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill uses shell-based
gitcommands (e.g.,git log,git shortlog,git dirstat) to perform repository analysis. These commands are used legitimately to generate metadata reports and orient the user in a new codebase. - [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from the repository's git history (commit messages and author names). Maliciously crafted commit messages could attempt to influence the agent's audit findings or subsequent recommendations.
- Ingestion points: Commit messages and history are read in
references/bug-clusters.mdandreferences/firefighting-signals.mdusinggit log. - Boundary markers: No explicit delimiters or instructions are provided to the agent to treat commit messages as untrusted content that might contain instructions.
- Capability inventory: The agent can execute shell commands (git) and write a report file (
git-codebase-preflight-report.md) to the local filesystem. - Sanitization: The commit messages are filtered by keywords (e.g., 'fix', 'bug') but are not otherwise sanitized or escaped before being presented to the agent for interpretation.
Audit Metadata