gh-stack
Pass
Audited by Gen Agent Trust Hub on Nov 2, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill manages the installation of the 'github/gh-stack' extension. This download originates from an official GitHub repository, which is a trusted source. The installation process is gated by a mandatory user approval step and is verified after execution.- [COMMAND_EXECUTION]: The skill executes several 'git' and 'gh' CLI commands to perform branch management and PR synchronization. These operations are protected by environment variable scrubbing (unsetting BASH_ENV and others) and require explicit per-use approval via the 'AskUserQuestion' tool. It also implements unshared attestation checks to ensure no other collaborators are working on the branches before force-pushing.- [INDIRECT_PROMPT_INJECTION]: The skill processes pull request titles and bodies which are external data sources. It mitigates this risk by implementing an attribution verification cycle that scans for leaks and applies remediation scripts to sanitize PR content before publication.
- Ingestion points: Pull request metadata (titles and bodies) retrieved via 'gh pr view'.
- Boundary markers: Not explicitly defined as delimiters, but content is treated as untrusted and passed through a dedicated sanitization script.
- Capability inventory: Subprocess execution of 'gh stack' and 'gh pr edit' across the main logic.
- Sanitization: Employs 'scripts/sanitize-pr-content.sh' and 'scripts/commit-msg-guard.sh' to validate and clean PR content before use.
Audit Metadata