gaip-broker
Click on "Deploy Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type
@followed by the MCP server name and your instructions, e.g., "@gaip-brokercan you check if this agent card is conformant? https://agent.example/.well-known/agent-card.json"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
GAIP
GAIP, the Governed Agentic Intelligence Platform, keeps an independent, dated record of what AI agents, MCP servers and online shops publish, and answers questions about that record with signed receipts. Free, read-only, no key; facts at a stated time, never an assessment.
First call: gaip_check with a url on https://www.gaipagents.com/mcp. This one runs as-is; then use the URL
of the agent, MCP server or API you care about:
curl -s https://www.gaipagents.com/mcp \
-H 'Content-Type: application/json' -H 'Accept: application/json, text/event-stream' \
-d '{"jsonrpc":"2.0","id":1,"method":"tools/call","params":{"name":"gaip_check","arguments":{"url":"https://mcp.deepwiki.com/mcp"}}}'Without MCP: GET https://www.gaipagents.com/v1/free/conformance/check?url=<url>.
The map: every question GAIP answers, grouped by who asks, with its tool and HTTP route: https://www.gaipagents.com/v1/free/map. Every completed answer carries a continuity handle; send it back and the answer is bound to your own history with GAIP.
One line to connect
Where | One line |
Any MCP client |
|
Claude Code |
|
Claude Desktop / claude.ai | Customize > Connectors > Add custom connector, URL |
Agent Skills (Claude Code, Cursor, Codex and others) |
|
Cursor, VS Code | the buttons below, or |
Gemini CLI |
|
OpenAI Agents SDK |
|
Vercel AI SDK |
|
LangChain |
|
CrewAI |
|
A2A (1.0 and 0.3.0) | agent card |
x402 | nothing to set up: GAIP is free and never answers with HTTP 402 ( |
Windsurf |
|
Cline |
|
Everything else, with copy-paste code |
Related MCP server: turva-mcp
In CI: two GitHub Actions
GAIP Watch fails your build when an MCP server or agent your agent depends on changed its tools, schemas, sign-in or prices, with dated receipts (watch/):
- uses: tjcgraham-rgb/gaip-agents/watch@main
with:
urls: https://mcp.example.com/mcpGAIP Agents check runs gaip_check on your own agent card, MCP server or OpenAPI document on every push:
- uses: tjcgraham-rgb/gaip-agents@v1
with:
agent-url: https://your-agent.example/.well-known/agent-card.json
fail-on-error: "true"Outputs: verdict (READY, FIXES_NEEDED, UNREACHABLE or NOT_CHECKED), receipt-id, error-count, warning-count.
Providers can show their own dated record with the badge
https://www.gaipagents.com/v1/free/watch/badge.svg?url=<your endpoint>.
In this repository
Path | What it is |
| The Official MCP Registry records: |
| The sentence, the first call, the map's rows, endpoints and example calls for directories (also served at |
| GAIP as an Agent Skill (agentskills.io format; also served at |
| Claude Code plugin and marketplace manifests: the skill plus the remote MCP server |
| Cursor plugin manifest and its MCP server |
| Gemini CLI extension: installs GAIP's MCP server and tells the model when to use each tool |
| Python client: |
| The check Action ( |
| The GAIP Watch Action |
| Dependency-free parser and test vectors for AI-preference signals (IETF AI Preferences |
| GAIP logo (SVG, and a 400×400 PNG) and the specialist icons (SVG) |
| Install steps for AI assistants that set up MCP servers themselves (Cline) |
The registry records, listing kit, packages, Watch Action, skill and Gemini extension are copies of what GAIP's service generates; a check in the service's repository warns when a copy here falls out of step. The service itself runs at www.gaipagents.com.
What an answer means
Every answer states what was observed and when. It is not a certification, endorsement, ranking or universal score, and not legal, financial or insurance advice. Receipts are signed with GAIP's Ed25519 key, published at https://www.gaipagents.com/.well-known/gaip-receipt-keys.json.
Privacy: https://www.gaipagents.com/privacy
Corrections and opt-out: https://www.gaipagents.com/corrections
Methodology: https://www.gaipagents.com/methodology
Security contact: https://www.gaipagents.com/.well-known/security.txt
Licence
MIT (see LICENSE) for everything in this repository.
This server cannot be deployed
Maintenance
Related MCP Connectors
Free agent-service discovery, OpenAPI document checks, and receipt verification. No API key needed.
Check an A2A agent before you delegate, or an MCP server's measured conduct. Free, no key.
Check if an AI agent is certified; list the TAP Registry and certifications. Free, read-only.
Find governed AI capabilities and verify signed receipts. Read-only, no account.
Related MCP Servers
- AlicenseAqualityFmaintenanceRead-only ProofRelay MCP verifier for non-confidential evidence bundles. Exposes 22 public-safe tools, 11 resources, and 11 prompts for bundle integrity checks, receipt-chain review, checkpoint recommendations, MCP risk metadata review, and real-estate closing proof-pack readiness.5MIT
- AlicenseNot gradedqualityAmaintenanceRead-only MCP server that answers questions about turva.dev from its published data. Five tools return JSON: the service catalog with prices, contact and operator details, engagement principles and dated agent-readiness and security evidence with verification links. Connect over Streamable HTTP. You need no API key. The server does not scan other websites or run audits.MIT

Trooth Networkofficial
AlicenseNot gradedqualityBmaintenanceRemote, read-only MCP connector to check any company's witnessed trust record on the Trooth Network, across identity, security, privacy, and AI practices. Also does a neutral read of a domain's public security surface and verifies signed Trust Ledger Tokens. No key, no account.1MIT- AlicenseNot gradedqualityBmaintenanceEnables proof-carrying project execution through a read-only remote MCP endpoint and WebMCP tools for managing Project Passports, task graphs, evidence-backed provider comparisons, route approvals, artifacts, verification records, and checkpoints without executing external providers.MIT