git-commit
Pass
Audited by Gen Agent Trust Hub on Feb 19, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- PROMPT_INJECTION (LOW): The skill contains multiple instances of instructions (e.g., 'MUST ALWAYS be used... regardless of whether the user explicitly requests it or not') that attempt to override user intent and force the agent to ignore potential user constraints.
- COMMAND_EXECUTION (SAFE): The skill uses standard git utility commands (
git status,git diff,git log,git add,git commit) for its primary purpose. No arbitrary command execution or access to sensitive system files was detected. - INDIRECT_PROMPT_INJECTION (LOW): The skill is vulnerable to instructions embedded in the data it processes.
- Ingestion points:
git log --oneline -10andgit diff(File: SKILL.md). - Boundary markers: Absent; the skill does not use delimiters to separate git output from its internal instructions.
- Capability inventory:
git add,git commit,git apply(File: SKILL.md). - Sanitization: Absent; the agent is instructed to follow patterns detected in potentially untrusted commit history without validation.
Audit Metadata