Documents are organized by reading path. Start with the path that matches your goal.
For hackathon judges and technical reviewers — understand the product boundary, then inspect the Reference Ops AI demo.
| Step | Document | Purpose |
|---|---|---|
| 1 | ../README.md |
Repo overview, quickstart, AgentGate vs Phoenix |
| 2 | PRD_PRODUCT.md |
Hackathon-scope product PRD and release checklist |
| 3 | REFERENCE_WORKFLOW.md |
Five-step Reference Ops AI demo walkthrough: homepage proof, blocked report, generated release controls, approved-with-warnings follow-up |
README → PRD_PRODUCT → REFERENCE_WORKFLOW
Optional supporting material: DEMO_CICD.md, getting-started/RELEASE_PIPELINE.md, ../examples/artifacts/README.md, getting-started/DEPLOYMENT.md.
For integrators attaching a production agent via AgentPack.
| Document | Purpose |
|---|---|
integration/README.md |
Integration hub and reading order |
integration/CONNECT_YOUR_AGENT.md |
Three-step entry |
integration/AGENTGATE_INTEGRATION_SKILL.md |
End-to-end integration guide |
integration/INTEGRATION_CONTRACT.md |
Certification tiers and span contract |
integration/AGENTPACK.md |
Suite and policy JSON |
integration/PHOENIX_INTEGRATION.md |
MCP, normalizer, eval labels |
integration/RELEASE_OUTPUT.md |
Audit bundle and Gemini boundaries |
Extended schema and gate rules: PRD.md (integrators and auditors).
| Document | Purpose |
|---|---|
REFERENCE_WORKFLOW.md |
Demo narrative and CLI |
integrations/stability_ops/README.md |
Canonical config paths and PII checklist |
Canonical configs: configs/agents/stability_ops/.
For contributors and AI coding agents — not required for hackathon demo.
| Document | Purpose |
|---|---|
../AGENTS.md |
Rules for AI agents in this repo |
../CONTEXT.md |
Domain vocabulary (AgentPack, EffectiveConfig, …) |
ARCHITECTURE.md |
System shape and modules |
adr/README.md |
Architecture decision records |
PRD.md |
Extended technical specification |