ship-fast
Pass
Audited by Gen Agent Trust Hub on Sep 9, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data, including Pull Request (PR) content and project instructions, while possessing the capability to execute commands and merge code. An attacker could potentially embed instructions in a PR to influence the agent during automated checks or merges. The skill mitigates this by requiring explicit human approval for stable releases and instructing the agent to avoid granting privileges to untrusted code.\n
- Ingestion points: Open Pull Requests, Git repository state, and project documentation (SKILL.md).\n
- Capability inventory: Executing project-defined check commands, merging PRs, and triggering deployment workflows (SKILL.md).\n
- Boundary markers: Instructions direct the creation of a project policy (references/project-policy.md) to define environment and authorization boundaries.\n
- Sanitization: There is no mention of explicit sanitization or filtering of data ingested from PRs.\n- [COMMAND_EXECUTION]: To function, the skill identifies and executes 'required check commands' and 'verification operations' defined in the project's environment. While necessary for the skill's purpose of CI/CD management, this involves running arbitrary commands provided by the project configuration (SKILL.md).
Audit Metadata