commit-submit
Pass
Audited by Gen Agent Trust Hub on Oct 9, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted user input (descriptions of code changes) and incorporates them into a structured format for use with version control tools. This creates an indirect prompt injection surface where instructions could be embedded in the commit text. However, the skill explicitly mandates that the agent must obtain user confirmation before executing a commit, which serves as a critical human-in-the-loop safety control.
- Ingestion points: Processes user-provided descriptions, issue numbers, and log messages in
SKILL.md. - Boundary markers: No specific delimiters or escaping instructions are provided for the user-supplied content within the commit template.
- Capability inventory: The skill is intended to be used with git tools for staging and committing changes.
- Sanitization: The skill includes a procedural safeguard requiring explicit user confirmation before any action is taken.
Audit Metadata