059-design-atdd
Pass
Audited by Gen Agent Trust Hub on Aug 3, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill is purely instructional and focused on documentation analysis. No malicious patterns, obfuscation, or unauthorized access attempts were found.
- [PROMPT_INJECTION]: The skill employs defensive instructions by mandating that acceptance criteria and test-like text be treated strictly as 'requirement data' rather than executable instructions, effectively mitigating indirect prompt injection risks. 1. Ingestion points: OpenSpec proposal, requirements, scenarios, and tasks.md checklist. 2. Boundary markers: Explicit instructions to treat test-like text as requirement data and never as executable instructions. 3. Capability inventory: The skill does not request access to shell execution or external network tools. 4. Sanitization: Instructs the agent to report source conflicts instead of resolving them silently.
- [DATA_EXFILTRATION]: The skill only accesses project-specific files necessary for its stated purpose and does not engage in external network communication or sensitive system file access.
Audit Metadata