workflow-spec-to-mockup

Pass

Audited by Gen Agent Trust Hub on Nov 9, 2026

Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill instructs the agent to run a local Node.js script for workflow maintenance: node .claude/scripts/lib/workflow-skill-contract.cjs --write.
  • [DYNAMIC_EXECUTION]: The skill implements a dynamic orchestration layer that calls other local sub-skills (e.g., $design-spec, $work-item, $ui-design) based on definitions within a workflow registry (workflows.json).
  • [INDIRECT_PROMPT_INJECTION]: The skill processes potentially untrusted external data from project specifications and configuration files.
  • Ingestion points: Ingests data from workflows.json, docs/project-config.json, and user-provided 'canonical specs' (e.g., docs/specs).
  • Boundary markers: The workflow relies on artifact profiles and registry-owned roles to structure data, though it lacks explicit 'ignore instructions' delimiters for the spec content itself.
  • Capability inventory: Ability to execute shell commands (node), spawn subagents (spawn_agent), and perform extensive file system writes in tmp/design/ and team-artifacts/.
  • Sanitization: Implements rigorous quality gates including design spec reviews ($work-item --mode=review) and UI walkthroughs ($ui-design --mode=review) to validate and fix outputs before completion.
Audit Metadata
Risk Level
SAFE
Analyzed
Nov 9, 2026, 01:34 AM