commit-message
Pass
Audited by Gen Agent Trust Hub on Sep 23, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill uses git and gh CLI tools to gather repository context and perform local commits. All repository modifications are gated by hard rules requiring explicit user approval via AskUserQuestion.
- [INDIRECT_PROMPT_INJECTION]: The skill reads data from local git diffs and logs which are technically untrusted inputs. This risk is managed through a mandatory verification step where the user must review the generated commit message in the chat interface before it is executed.
- Ingestion points: git diff --cached and git log (SKILL.md).
- Boundary markers: Explicit confirmation prompts using AskUserQuestion.
- Capability inventory: git commit, git add.
- Sanitization: Relies on human-in-the-loop review of generated output.
Audit Metadata