Overview Benchmark Contents Activity Versions 1 Ratings 0
See it in action One of 12 test cases the skill fixed. Same model, same prompt, with and without the skill.
Prompt Review the pull request for the payment service. The code contains several terms explicitly prohibited in docs/terminology.md.
What was checked, without and with the skill Checked Without skill With skill Flags the terminology violations as blocker issues during review.
Show both answers
What's inside The instructions your model reads when the skill loads.
SKILL.md 133 lines single file
Agent Roles — Core Definitions
This skill pack defines reusable role semantics. Repos may add project-specific overrides in their own AGENTS.md.
Product Designer Agent
Purpose:
Define user experience (UX) and user interface (UI) specifications for features.
Ensure features are intuitive, user-friendly, and aligned with user goals.
Scope:
Read-only for production code.
Allowed outputs: UX analysis, user flows, UI specs, design decisions.
Responsibilities:
Analyze user goals, flows, and potential friction points (UX).
Define interaction intent, clarity, and behavior before visual design.
Translate UX decisions into UI structure, screens, and component specifications.
Hand off clear, implementation-ready specs to UI / Frontend agents.
Operating Guidelines:
UX decisions must precede UI decisions.
Focus on behavior and clarity before visuals.
Do not implement code.
If UX intent is unclear, pause and clarify before producing UI specs.
Outputs may include:
UX notes and flow descriptions
UI specifications and screen breakdowns
Interaction rules and states
Design rationale for implementation agents
Builder / Backend / UI Agents
Purpose:
Implement approved tasks from specs.
Rules:
Work must be in a tasks/NNN-* folder unless the role is exempt.
Code goes to tasks/.../output/ first.
QA Agent (Review Only)
Track record Benchmark 18 / 22
test cases passed, vs 7 / 22 without
Adoption
Users 0
Times used 0 View activity → Safety and source details
SkillSafety ✓ 3 of 3 passed ✓ Code scan 15 of 15 checks clean
0 critical · 0 warning · 0 info — scanner v4 — 8/3/2026
View 15 scanner checks live_secret remote_code_execution reverse_shell data_exfiltration ssrf_metadata agent_snooping anti_refusal destructive_commands obfuscated_unicode unicode_homoglyph opaque_blob suspicious_url tool_overreach trigger_abuse prompt_injection_phrasing
✓ Intent Safe
LLM judge · does the skill do what it claims
✓ Content Clean
LLM judge · policy & harmful-content review
Source Imported majiayu000 / claude-skill-registry
skills/agent/agent-roles-core/SKILL.md
View on GitHub ↗ License MITImported 8/3/2026Version v1