dayan-agent-skills

Artifact fixtures

These fixtures show what useful Skill output can look like after a clean run. They are not transcripts, benchmark claims, customer work, or model output guarantees.

Each fixture includes:

• Skill: the public Skill being demonstrated;
• Input: the safe public request shape;
• Output artifact: the copyable result structure;
• Verification: how a reviewer can check the result;
• Boundary: what the fixture does not prove.

Fixtures

• Repository orientation map — dayan-orient
• Support triage agent contract — dayan-agent-designer
• Public redline hook spec — dayan-hook-factory
• Pagination API review — dayan-api-design-reviewer
• AI visibility docs audit — dayan-ai-seo
• Visual deck brief — dayan-deck
• Architecture decision record — dayan-tech-stack-evaluator
• Agent package manifest — dayan-agent-factory
• Responsive UI acceptance plan — dayan-frontend-responsive-ui
• Tool-before-agent verifier plan — dayan-hook-factory
• False-positive control review — dayan-adversarial-reviewer
• Human authority ledger — dayan-agent-designer

Add a fixture

Use the same five-section shape and keep it public:

1. no private paths;
2. no secrets or credentials;
3. no real customer identifiers;
4. no claims that the fixture proves production success;
5. one clear verification command or checklist.

Then run:

python3 scripts/verify_fixtures.py

This site is open source. Improve this page.