writing-commit-messages
Pass
Audited by Gen Agent Trust Hub on Jul 20, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill instructs the agent to run local CLI commands (git diff/commit or jj diff/commit) to manage repository state.
- [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection from untrusted data.
- Ingestion points: The skill reads output from git diff (untrusted data).
- Boundary markers: Absent; no delimiters are used to separate the diff data from the agent's instructions.
- Capability inventory: File system read (diff) and write (commit) via shell commands.
- Sanitization: Absent; the skill does not include steps to sanitize the diff output before processing.
Audit Metadata