brainstorming
Pass
Audited by Gen Agent Trust Hub on Feb 17, 2026
Risk Level: SAFE
Full Analysis
- [Prompt Injection] (SAFE): The instructions focus on structured design phases and do not contain patterns for bypassing safety filters, extracting system prompts, or ignoring instructions.
- [Data Exposure & Exfiltration] (SAFE): No hardcoded credentials, sensitive file path access, or network calls (like curl or wget) were detected.
- [Unverifiable Dependencies & Remote Code Execution] (SAFE): The skill references sub-skills like 'superpowers:using-git-worktrees', which is a standard pattern for modular agent capabilities. There are no external package installations or remote script executions.
- [Indirect Prompt Injection] (LOW): The skill processes user input to create designs. Evidence: 1. Ingestion points: Phase 1 (Understanding) and Phase 3 (Validation). 2. Boundary markers: Absent. 3. Capability inventory: File writing to 'docs/plans/' and execution of git worktree/planning sub-skills. 4. Sanitization: Absent. The risk is considered low as the context is limited to design documentation and development workflow.
- [Obfuscation] (SAFE): The content is clear markdown with no Base64, zero-width characters, or homoglyphs.
Audit Metadata