design-spec

Pass

Audited by Gen Agent Trust Hub on Nov 10, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill has an attack surface for indirect prompt injection as it processes untrusted user data (wireframes, screenshots, task descriptions, and design tool links) to generate instructions for implementation.
  • Ingestion points: Workflow Step 1 in SKILL.md and 'Input Routing' in references/wireframe.md identify external images and text as the primary inputs for specification generation.
  • Boundary markers: The protocol requires all claims to be explicitly tagged as SOURCED or INFERRED and mandates the creation of a 'Journey Report' (UX-1) to verify intent and context before any specification is authored.
  • Capability inventory: The skill is authorized to read project configurations (docs/project-config.json) and write design artifacts (design-specs/*.md) to the file system, as well as update frontmatter in feature specifications.
  • Sanitization: The skill enforces 'M1-M7 Compliance' which requires the output to be tech-agnostic and described in plain UX language. This requirement effectively prevents the direct pass-through of malicious code, framework identifiers, or technical commands from the input into the resulting specification prose.
Audit Metadata
Risk Level
SAFE
Analyzed
Nov 10, 2026, 08:16 PM