AGENTIFY
You've built something. Help the agents that are looking for it find their way in.
the promise
out of 100 · average website 2026 · cloudflare agent-readiness scanner
Across Cloudflare's April 2026 scan of 62 domains, no site scored above 70.
Run this skill on any repo. It audits 30 points. Scaffolds the missing layers. Verifies the deploy live.
Two to seven days. 25-30 of 30. Higher than every site Cloudflare scanned.
Request access.
Private beta. Each request is reviewed by Ray personally. You'll get a GitHub repo invite if approved.
Or, book a session.
Live walk-through with Ray. Audit your stack together. Leave with a roadmap.
Install.
Three commands once you've been added. That's the whole thing.
-
Clone
Clones the skill into your Claude skills directory.
git clone https://github.com/RayyanZahid/agentify ~/.claude/skills/agentify -
Restart
Reload Claude Code. The skill auto-registers.
claude --restart -
Invoke
Type this in any session inside a repo.
/agentify
Prompts.
Tap any to copy. Paste into Claude Code, Cursor, or any agent that loads Anthropic Skills.
Audit ./my-repo for agent-readiness. Score against the 30-point checklist and list the top three gaps.
I am launching a new SaaS. Make it agent-first from day one. Scaffold all five layers using lean managed defaults. Atomic commit per layer.
Add the well-known agent files to my Next.js app. agent-card.json, mcp.json, ai-agent.json, llms.txt, agent-aware robots.txt. CORS open on /.well-known/.
Wire RFC 8628 device-code agent signup into my Clerk-protected Next.js app. Three endpoints plus the state machine. See agentify reference 21.
Extend my agent surface to 25 of 30. Add signed requests per RFC 9421, a per-user activity log, an agent feedback endpoint, a recovery block in ai-agent.json, and the internal-mint endpoint with audit log + scope ceiling.
Scaffold the AET (Agent Experience Testing) framework into my repo. Eight sample scenarios including cross-surface diff, refusal-path, fixture-leak, and manifest-signature consistency. Wire the GitHub Actions smoke job with an internally-minted token.
Run the live verification script on my-domain.com. Confirm content-shape, behavioral MCP + A2A probes, and refusal paths all pass. Then run agentify freshness --py --append data/freshness-timeseries.jsonl to log drift.
Ship it. Then prove it works.
You can post the manifest. Wire MCP. Mount the A2A endpoint. None of that proves an agent can actually use your surface.
The Agent Experience Testing dimension is five checklist points. The same harness ran against immersive-commons-unified on 2026-05-17 and caught nine real bugs in one push. A four-agent team fixed all nine; scenario coverage went from 22 of 37 tools to all 37.
The skill ships eight sample scenarios. A runner. Four drivers. A GitHub Action that walks the surface on every push with an internally-minted token. No human in the loop.
What a run looks like.
Sample of the agentify audit . output format. Real structure. Per-point passes and fails vary by repo. A live audit of immersive-commons-unified on 2026-05-17 returned 20/30. Production-ready, with four named cross-cutting gaps the team accepted at launch.
# Agent-readiness audit **Repo:** ./my-saas **Date:** 2026-05-17T00:00:00Z ## Layer 1 — Discovery (4) ✓ 1. /.well-known/agent-card.json ✓ 2. /.well-known/mcp.json ✓ 3. /.well-known/ai-agent.json (Aiia) ✓ 4. /llms.txt + agent-aware robots.txt ## Layer 4 — Execution (4) ✓ 10. MCP server present ✓ 11. SKILL.md (Anthropic Skill) ✗ 12. Accessibility tree (found <div onclick> on /dashboard) ✓ 13. WebMCP exposure ## Cross-cutting (6) ✓ 16. Observability wired (langfuse) ✓ 17. Memory layer (mem0) ✓ 18. Durable execution (inngest) ✓ 19. Tool auth (composio) ✓ 20. Agent CI/CD ✓ 21. Live verification (CF scanner) ## Layer 6 — Verification (5) — AET dimension ✓ 26. AET scenario suite present (5 scenarios) ✓ 27. Cross-surface consistency (N-way diff) ✓ 28. Refusal-path for protected endpoints ✗ 29. Cleanup-aware write fixtures ✓ 30. Per-tool latency budgets ## Score: 27/30 SOTA agent surface. Top one percent of the web. Next: wire cleanup hooks on write fixtures · point 29.
the spec · 30 points · 8 layers
checkpoints · grouped by layer · one atomic commit each
license
PROPRIETARY
© 2026 Rayyan Zahid. All rights reserved.
Not open source. No use, copying, redistribution, or derivative works without a written license. Access is granted per-recipient on request and may be revoked.