git-atomic-commit
Pass
Audited by Gen Agent Trust Hub on Feb 27, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: No malicious patterns or high-risk vulnerabilities were detected. The skill uses standard Git commands and a provided local bash script for commit management.
- [PROMPT_INJECTION]: While the skill reads untrusted data from file diffs, which is a standard indirect prompt injection surface, it does not possess network access or high-privilege capabilities that would enable exploitation. Ingestion points: File content and metadata read via Git diff commands. Boundary markers: None provided in the instructions. Capability inventory: Limited to Git commands and local script execution. Sanitization: Relies on the agent's underlying safety filters and the mechanical nature of the task.
Audit Metadata