skills/chann/skills/git-commit-push/Gen Agent Trust Hub

git-commit-push

Pass

Audited by Gen Agent Trust Hub on Sep 15, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill implements restricted Git workflows with clear security constraints against destructive operations like force pushing.
  • [DATA_EXPOSURE]: It includes a robust 'Red Flags' list that prevents the inclusion of sensitive files such as .env, .pem, .key, and SSH credentials in commits.
  • [COMMAND_EXECUTION]: Command execution is limited to standard Git CLI operations with no evidence of unsafe shell interpolation or remote script execution.
  • [INDIRECT_PROMPT_INJECTION]: 1. Ingestion points: reads file content and directory structures for staging; 2. Boundary markers: references a parent skill for workflow and markers; 3. Capability inventory: git commit, git push, and git branch; 4. Sanitization: relies on the standard git-commit workflow for review before execution.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 15, 2026, 04:13 PM