git-governance
Pass
Audited by Gen Agent Trust Hub on Nov 5, 2026
Risk Level: SAFECOMMAND_EXECUTIONPRIVILEGE_ESCALATIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill utilizes shell commands for repository management (git) and GitHub integration (gh), as well as executing a local Gradle wrapper (./gradlew) for testing.
- [PRIVILEGE_ESCALATION]: The instructions explicitly direct the agent to use 'BypassSandbox: true' to perform network-reliant tasks and access system credentials, effectively instructing the agent to operate outside its default security boundaries.
- [INDIRECT_PROMPT_INJECTION]: The skill processes potentially untrusted data from GitHub and local files, creating a surface for injection attacks.
- Ingestion points: Data is ingested from 'gh issue list', 'gh pr list', and project tracking files ('feature_list.json', 'PROGRESS.md').
- Boundary markers: No specific boundary markers or instructions are provided to the agent to handle untrusted content within fetched GitHub issues or PRs.
- Capability inventory: The agent possesses capabilities to execute code, modify the file system, and interact with remote repositories.
- Sanitization: There are no defined processes for sanitizing or validating external data before it influences the agent's actions or output.
Audit Metadata