openspec-onboard
Pass
Audited by Gen Agent Trust Hub on Sep 27, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill uses local CLI tools including 'openspec' and 'git' to verify installation, analyze code history, and manage the project workflow. These commands are restricted to the local development environment and are expected for the skill's purpose.
- [INDIRECT_PROMPT_INJECTION]: The skill scans codebase files for comments such as TODO or FIXME and recent git logs to suggest tasks. This introduces a surface for untrusted data to enter the agent context. However, the risk is minimized by the skill's design, which requires user confirmation for task selection and artifact drafting, preventing autonomous execution of potentially malicious code comments.
Audit Metadata