MCP authorization: Authorization (draft)
Provision mcp.02 · Model Context Protocol, Authorization (draft) · requirement
MCP authorization servers MUST provide at least one discovery mechanism: OAuth 2.0 Authorization Server Metadata (RFC 8414) or OpenID Connect Discovery 1.0.
The text is as taken from the source. In the source: https://modelcontextprotocol.io/specification/draft/basic/authorization#overview (the section of the specification that holds the requirement (its opening words were found there); linked 2026-11-10). The instrument: Model Context Protocol, Authorization (draft), the project's; requirements quoted with the source; read 2026-11-09.
Who cites it
| Indicator | The fact it asks for | Kind of link | The assertion |
|---|---|---|---|
I24 | Its authorisation is published as metadata that anyone can read | same: evidenced | Asserted in the mapping by gamma-ray@web-summit.dinis; the written basis is not recorded yet. |
S3 | It is machine-readable for agents: it publishes an llms.txt, an AGENTS.md, an API description or an agent card | part: partly evidenced | The requirement is a discovery mechanism; a machine-readable index lets another agent discover what is there. Asserted by gamma-ray@web-summit.dinis, 2026-11-10; weak: one reader, unchecked. |
A link is our assertion, not the instrument's. "Same" carries the evidence to the provision in full. "Part" marks the provision as partly evidenced: the fact is one piece of what it asks. "Negation" points the other way.
The evidence that reaches it
| Card | Through | This provision is | The evidence | How far to trust it |
|---|---|---|---|---|
| RiskMandate Publisher | S3 | partly evidenced | Source | π Moderate · S1·R5·A1 ×2 |
| The PrompTugal.news agent team | S3 | partly evidenced | Source | π Moderate · S1·R5·A1 ×2 |
| AI Reviewer | S3 | partly evidenced | "Key facts for accurate citation:" Source | π Moderate · S1·R5·A1 ×3 |
| ByTheLaw | S3 | partly evidenced | "The product itself (sign-up and the web app) lives on a separate domain" Source | π Moderate · S1·R5·A1 ×3 |
| Claude Code on the web | S3 | partly evidenced | Source | π Moderate · S1·R5·A1 ×3 |
| Claude Code, on your own machine | S3 | partly evidenced | Source | π Moderate · S1·R5·A1 ×3 |
| GetVocal AI agents | S3 | partly evidenced | "GetVocal builds AI voice and chat agents for regulated, high-volume customer operations." Source | π Moderate · S1·R5·A1 ×3 |
| GitHub Copilot cloud agent | S3 | partly evidenced | Source | π Moderate · S1·R5·A1 ×3 |
| Hackian | S3 | partly evidenced | "Ethiack REST API documentation." Source | π Moderate · S1·R5·A1 ×3 |
| Junie | S3 | partly evidenced | "Machine-readable endpoints" Source | π Moderate · S1·R5·A1 ×3 |
| Manus | S3 | partly evidenced | "Manus Documentation" Source | π Moderate · S1·R5·A1 ×3 |
| Noxus AI Workers | S3 | partly evidenced | ""title":"Noxus Backend"" Source | π Moderate · S1·R5·A1 ×3 |
| OutSystems Agent Workbench | S3 | partly evidenced | "OutSystems combines low-code simplicity with AI software development" Source | π Moderate · S1·R5·A1 ×3 |
| Replit Agent | S3 | partly evidenced | "Features (241 pages)" Source | π Moderate · S1·R5·A1 ×3 |
| Spinnable AI employees | S3 | partly evidenced | "AI Workforce Fit Guide for AI Agents" Source | π Moderate · S1·R5·A1 ×3 |
| Talkdesk Autopilot | S3 | partly evidenced | "# Purpose: AI usage guidance for Large Language Models (LLMs)" Source | π Moderate · S1·R5·A1 ×3 |
| Visor.ai AI agents | S3 | partly evidenced | "Build AI Agents that learn from your business" Source | π Moderate · S1·R5·A1 ×3 |
Touched never means complies.
Answer it directly
A company can answer this line item itself, in its own file at /.well-known/fact-ray.json, with the node gamma-ray:provision/mcp.02 (how). A company that already assesses itself against MCP authorization can reuse that work.