cspec-blueprint
Pass
Audited by Gen Agent Trust Hub on Mar 18, 2026
Risk Level: SAFE
Full Analysis
- [PROMPT_INJECTION]: The skill instructions do not contain attempts to override safety guidelines or hijack agent behavior.
- [DATA_EXFILTRATION]: The skill restricts its operations to the local .cspec directory and does not contain any network communication or credential harvesting logic.
- [EXTERNAL_DOWNLOADS]: No external dependencies, packages, or remote scripts are referenced or executed.
- [INDIRECT_PROMPT_INJECTION]: The skill has a surface for indirect prompt injection through its ingestion of untrusted specification data, but it is considered safe as this is its primary intended use. 1. Ingestion points: Reads architecture and slice specifications from the local .cspec directory and user-provided files. 2. Boundary markers: No explicit delimiters are used to wrap ingested data. 3. Capability inventory: Generates and writes markdown documentation to the .cspec/plans/ directory. 4. Sanitization: The skill does not perform specific sanitization or filtering of the input specification content.
Audit Metadata