Global AI Governance MCP server
Server Configuration
Describes the environment variables required to run the server.
| Name | Required | Description | Default |
|---|---|---|---|
| HOST | No | Bind address. | 127.0.0.1 |
| PORT | No | Bind port. | 8000 |
| GAG_AGENTIC_PATH | No | Use other checkouts instead of vendor/ for the agentic-ai-governance. | |
| GAG_TOOLKIT_PATH | No | Use other checkouts instead of vendor/ for the global-ai-governance-toolkit. | |
| GAG_MCP_BEARER_TOKEN | No | Optional static bearer token for private pilots. This is not OAuth. | |
| GAG_MCP_CORS_ORIGINS | No | Comma-separated origins for browser-based clients (e.g., MCP Inspector). Off by default. | |
| GAG_MCP_ALLOWED_HOSTS | No | Comma-separated Host allow-list (DNS-rebinding protection). Set this on any public host. |
Instructions
Guidance the server publishes about itself, which clients place ahead of the tool catalog so the model reads it before choosing anything.
This server publishes no instructions, or was last inspected before Glama recorded them.
Capabilities
Features and capabilities supported by this server
Protocol revision2025-11-25
| Capability | Details |
|---|---|
| tools | {
"listChanged": false
} |
| prompts | {
"listChanged": false
} |
| resources | {
"subscribe": false,
"listChanged": false
} |
Tools
Functions exposed to the LLM to take actions
| Name | Description |
|---|---|
| check_inventoryA | Run the Global AI Governance toolkit's policy-as-code checks on an AI system inventory CSV (toolkit template columns) and return findings by severity. Set use_sample=true to run the synthetic sample. Output is non-authorizing (authority_effect NONE). |
| build_decision_packA | Run the toolkit pipeline and return the Decision Pack manifest (files + SHA-256 digests), the integrity-check result, and the reference-only handoff. include_documents=true also returns the Markdown documents for the decision-maker. The pack prepares a decision; it is not the decision. |
| validate_passportA | Verify an Agent Trust Passport (agentic-ai-governance reference validator): schema, semantics, Ed25519 signature, key trust, validity window, and signed revocation. Use example='signed-unrevoked' or 'signed-revoked' for synthetic demos, or pass passport_json and trusted_key_json (and optionally revocation_list_json) as JSON strings. Verification is not authorization (authority_effect NONE). |
Prompts
Interactive templates invoked by user choice
| Name | Description |
|---|---|
No prompts | |
Resources
Contextual data attached and managed by the client
| Name | Description |
|---|---|
No resources | |
TDQS
Scored across 3 tools
Each tool targets a distinct artifact and action: checking an AI inventory CSV, building a Decision Pack, and verifying an Agent Trust Passport. There is no meaningful overlap in purpose or inputs, so an agent can select correctly without hesitation.
All three names follow a clean snake_case verb_noun pattern (check_inventory, build_decision_pack, validate_passport). The convention is applied uniformly with no style drift.
Three tools is on the lean side for a server with a broad 'Global AI Governance' name, but each tool represents a substantial, self-contained workflow. It is slightly under-scoped rather than padded, which is acceptable.
The surface covers the core verification/assessment workflows (inventory policy checks, decision packaging, passport validation) with clear non-authorizing semantics. It lacks any authoring or lifecycle operations (e.g., generating inventories or issuing passports), but those may fall outside the server's reference-toolkit scope.