checkpoint-plan
Pass
Audited by Gen Agent Trust Hub on Nov 17, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill analyzes implementation plans and repository configuration files to identify build and test commands.
- Ingestion points: Processes user-provided plans and local codebase metadata files (referenced in SKILL.md).
- Boundary markers: The skill incorporates strong procedural boundaries by requiring explicit user approval of the plan and every checkpoint.
- Capability inventory: File system reading, command execution for testing, and code modification (SKILL.md).
- Sanitization: Commands are extracted from the repository configuration and presented to the user for review within the plan template.
- [SAFE]: The skill implements significant security safeguards by enforcing a human-in-the-loop workflow, explicitly prohibiting autonomous commits, pushes, or PR creations.
Audit Metadata