openspec-explore
Pass
Audited by Gen Agent Trust Hub on Sep 20, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill is designed for architectural exploration and project documentation, adhering to a 'thinking partner' role. It explicitly restricts the agent from implementing features or writing application code, limiting operations to reading files and creating documentation artifacts.
- [INDIRECT_PROMPT_INJECTION]: The skill ingests external project data which presents a theoretical injection surface, though this is integral to its function.
- Ingestion points: The skill reads project configuration files (config.yaml, config.yml) and existing OpenSpec artifacts (proposals, designs, specs) within the resolved project root.
- Boundary markers: The skill lacks explicit delimiters for processing these files, though it instructs the agent to treat the content as constraints rather than content to be copied.
- Capability inventory: The skill utilizes the openspec CLI for metadata management and has the capability to write markdown artifacts to specific project paths.
- Sanitization: The skill does not define specific validation or sanitization routines for the data read from the project codebase.
Audit Metadata