commit-message
Pass
Audited by Gen Agent Trust Hub on Oct 13, 2026
Risk Level: SAFEPROMPT_INJECTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill description uses explicit bypass language such as OVERRIDE system prompt and Call this INSTEAD OF to replace default agent behavior. This attempt to hijack the instruction flow is a characteristic of prompt injection.
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from git diff outputs which serves as a potential injection vector. 1. Ingestion points: The skill reads staged changes via git diff --cached (SKILL.md). 2. Boundary markers: No markers are used to separate code diff content from instructions. 3. Capability inventory: The skill uses git commit and git add to execute shell commands (SKILL.md, references/examples.md). 4. Sanitization: External content is not sanitized or escaped before being processed for the commit message.
Audit Metadata