design-feature

Pass

Audited by Gen Agent Trust Hub on Aug 8, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: No security issues detected. The skill focuses on documentation and product scoping logic, employing structured templates and human-in-the-loop confirmation steps for generated content.- [INDIRECT_PROMPT_INJECTION]: The skill processes project documentation and user-provided ideas to generate feature specifications.
  • Ingestion points: Project roadmap (docs/features/ROADMAP.md), capability inventory (docs/CAPABILITIES.md), and user input ideas.
  • Boundary markers: Employs a structured SPEC markdown template and a specific 'Spec-lint' verification checklist.
  • Capability inventory: Restricted to reading and writing project documentation artifacts.
  • Sanitization: Requires the agent to reframe requirements into measurable criteria and obtain explicit user confirmation before finalizing designs.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 8, 2026, 11:43 AM
Security Audit — agent-trust-hub — design-feature