git
Pass
Audited by Gen Agent Trust Hub on Sep 21, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill relies on extensive execution of command-line interface (CLI) tools including git, the GitHub CLI (gh), GitLab CLI (glab), and Forgejo CLI (fj).
- Evidence: The skill provides complex command sequences for managing branches, resolving conflicts, and automating releases across multiple remotes.
- Safeguards: Explicit instructions are provided to the agent to require user confirmation for all destructive operations, such as hard resets or force pushes.
- [EXTERNAL_DOWNLOADS]: The skill provides instructions for installing and utilizing various third-party security and productivity tools.
- Evidence: It references the installation of
gitsign,gitleaks,git-filter-repo, andforgejo-cli. - Verification: All external references target official package registries (PyPI, Go) or well-known, trusted organizations like Sigstore, the Forgejo community, and official GitHub repositories.
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data from repository logs and forge-specific interfaces (PR/MR descriptions).
- Ingestion Points: The skill uses
gh pr view,glab mr view, andgit logto gather context from external sources. - Boundary Markers: The skill implements a robust "AI Self-Check" and specific rules that instruct the agent to ignore AI-generated artifacts, respect repository-specific instruction files, and avoid unauthorized AI attribution in commit messages.
- Capability Inventory: The agent possesses broad system capabilities, including file modification and remote network communication, which are necessary for git operations but carefully scoped by the self-check guidelines.
- [SAFE]: The skill implements advanced security hardening and follows best practices for credential management.
- Credential Security: The skill provides secure patterns for handling API tokens, such as piping secrets to stdin or using mode-0600 configuration files, specifically to prevent sensitive values from appearing in command arguments or shell history.
- Security Awareness: It includes an up-to-date reference list of recent Git and forge vulnerabilities (2024–2026) to inform the agent and user of potential risks when cloning untrusted repositories.
Audit Metadata