GAIP Integration & Protocol Agent MCP server
io.github.tjcgraham-rgb/gaip-integration-protocoltjcgraham-rgbstreamable-httpv1.5.0Scanned
Behind Check and GAIP Watch, and Diagnose: A2A, MCP and OpenAPI checks with a dated change log.
13 tools: 1 can delete or cancel things, 1 do something that can't be undone, 9 change things and 2 only read.
- Reads private data · 1
- Reads text others write · 5
- Can send data out · 6
- 1 finding
Tools listed without sign-in. Remote posture
With Anlyon
1 tool here can delete or cancel things and 1 does something that can't be undone. Govern them.
- Anlyon lists the tools and adopts none.
- Each tool you adopt gets approvals and receipts.
- Your agent never holds the key.
A tool your agent calls directly is not routed through Anlyon and is not governed by it. Anlyon Radar only lists and labels, and these labels are the same either way.
What the GAIP Integration & Protocol Agent MCP server can do
The GAIP Integration & Protocol Agent MCP server is published by the GitHub account tjcgraham-rgb. It has 13 tools. It can delete or cancel things (1 tool: gaip_watch_remove), do things that cannot be undone, such as sending, paying or publishing (gaip_observatory_capture) and create or change data (9 tools: check_agent_conformance, gaip_watch_register, gaip_watch_events, gaip_manifest_changes and 5 more). 2 tools only read.
1 tool reads private data (gaip_compare_supplier_snapshots); 5 tools bring in text other people can write (check_agent_conformance, gaip_watch_register, gaip_observatory_lookup and 2 more); 6 tools can send data out of your system (check_agent_conformance, gaip_watch_register, gaip_observatory_lookup and 3 more). On its own it has all three parts of a prompt-injection exfiltration path: private data, text others write, and a way out.
The fixed checks matched 1 finding (annotation-conflict). Each one comes with the text it matched.
Its remote endpoint lists its tools without sign-in.
Last scanned Nov 21, 2026, version 1.5.0.
Tools
Grouped by what each tool does, most consequential first. Open one for its reasons and full description; descriptions are the publisher's own words.
13 tools
Deletes or cancels 1Claude Code: deny
gaip_watch_removeUse this when asked to stop watching an agent or service.No findings
Why it is in this group
- The name "gaip_watch_remove" says "remove": a word for removing or ending something.
- The server marks it destructive (destructiveHint: true).
Description, as the server wrote it Use this when asked to stop watching an agent or service. Pass your `continuity_handle` and the `watch_id`. Its stored events are deleted; retained receipts stay verifiable. Free, no account; records a GAIP receipt only. Inputs must be public and non-personal.
Annotations the server declares
- destructiveHint: true
- idempotentHint: true
- openWorldHint: false
- readOnlyHint: false
Can't be undone 1Claude Code: ask
gaip_observatory_captureUse this to save a dated record of what a public MCP server or A2A agent card declares now.Reads text others writeCan send data outNo findings
Why it is in this group
- The name "gaip_observatory_capture" says "capture": a word that usually means money moves, which cannot be taken back.
What it can reach
- untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.
- outboundThe input takes a URL (input.url): a request to a URL can carry data to whoever runs it.
Description, as the server wrote it Use this to save a dated record of what a public MCP server or A2A agent card declares now. Pass its `url`. GAIP reads it once (robots.txt respected, never tools/call), re-reads it daily and returns the capture id, time and SHA-256 hashes. Not an assessment. Free, no account; records a GAIP receipt only. Inputs must be public and non-personal.
Annotations the server declares
- destructiveHint: false
- idempotentHint: false
- openWorldHint: true
- readOnlyHint: false
Changes things 9Claude Code: ask
check_agent_conformanceUse this when asked to check an AI agent or MCP server: does it work, is it reachable, is its card valid ("check this agent", "is this MCP server up?").Reads text others writeCan send data outNo findings
Why it is in this group
- Nothing in its name or description says what it does, so it is counted as a write.
- The server marks it read-only (readOnlyHint: true), but an annotation never lowers a class.
What it can reach
- untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.
- outboundThe input takes a URL (input.url): a request to a URL can carry data to whoever runs it.
Description, as the server wrote it Use this when asked to check an AI agent or MCP server: does it work, is it reachable, is its card valid ("check this agent", "is this MCP server up?"). Pass just its `url`. Returns READY / FIXES_NEEDED / NO_AGENT_CARD / UNREACHABLE with fixes and a receipt. Free, read-only, no account; inputs must be public and non-personal.
Annotations the server declares
- destructiveHint: false
- idempotentHint: true
- openWorldHint: true
- readOnlyHint: true
gaip_check_permissionUse this before your agent fetches a page ("may my agent fetch this?").Reads text others writeCan send data outNo findings
Why it is in this group
- Nothing in its name or description says what it does, so it is counted as a write.
- The server marks it read-only (readOnlyHint: true), but an annotation never lowers a class.
What it can reach
- untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.
- outboundThe input takes a URL (input.url): a request to a URL can carry data to whoever runs it.
Description, as the server wrote it Use this before your agent fetches a page ("may my agent fetch this?"). Pass the `url` and your `agent` token. Reads robots.txt (RFC 9309), AI-preference lines, RSL, tdmrep.json and llms.txt; returns ALLOWED, DISALLOWED, NO_RULE or UNKNOWN, the rule quoted, the time and a signature. Free, read-only, no account; inputs must be public and non-personal.
Annotations the server declares
- destructiveHint: false
- idempotentHint: true
- openWorldHint: true
- readOnlyHint: true
gaip_compare_supplier_snapshotsUse this when asked what changed between two versions of an agent card, MCP server description or OpenAPI document ("did this API break?").Reads private data1 finding
Why it is in this group
- The name "gaip_compare_supplier_snapshots" says "compare": a word for reading.
- The server marks it as not read-only (readOnlyHint: false).
What it can reach
- private-dataThe description's first sentence says "document": it reads documents.
Findings
- Warn
The server marks "gaip_compare_supplier_snapshots" as not read-only (readOnlyHint: false), and its words say it reads.
What this check looks forMatched text
annotations: {"readOnlyHint":false,"destructiveHint":false,"idempotentHint":false,"openWorldHint":false}; The name "gaip_compare_supplier_snapshots" says "compare": a word for reading.
Description, as the server wrote it Use this when asked what changed between two versions of an agent card, MCP server description or OpenAPI document ("did this API break?"). Pass `prior` and `current`. Returns each change as BREAKING / NON_BREAKING / INFO with migration notes and a receipt. Free, no account; records a GAIP receipt only. Inputs must be public and non-personal.
Annotations the server declares
- destructiveHint: false
- idempotentHint: false
- openWorldHint: false
- readOnlyHint: false
gaip_diagnose_integration_failureUse this when an agent, MCP or API call failed ("why did my call fail?", "what does this 429 mean?").No findings
Why it is in this group
- Nothing in its name or description says what it does, so it is counted as a write.
Description, as the server wrote it Use this when an agent, MCP or API call failed ("why did my call fail?", "what does this 429 mean?"). Pass the `failure` ({code, detail}) or the raw error text. Returns the likely cause, when and how to retry, fix steps and a receipt. GAIP touches no other system. Free, no account; records a GAIP receipt only. Inputs must be public and non-personal.
Annotations the server declares
- destructiveHint: false
- idempotentHint: false
- openWorldHint: false
- readOnlyHint: false
gaip_manifest_changesUse this when asked which MCP servers or agents changed recently ("what changed this week?").No findings
Why it is in this group
- Nothing in its name or description says what it does, so it is counted as a write.
- The server marks it read-only (readOnlyHint: true), but an annotation never lowers a class.
Description, as the server wrote it Use this when asked which MCP servers or agents changed recently ("what changed this week?"). Optional `since` (RFC 3339), `target_id`, `limit` (max 100). Returns the Observatory's hash-chained log of manifest changes: before/after hashes, neutral facts, receipts. Free, read-only, no account; inputs must be public and non-personal.
Annotations the server declares
- destructiveHint: false
- idempotentHint: true
- openWorldHint: false
- readOnlyHint: true
gaip_supplier_reliabilityUse this when a supplier agent or API you depend on started failing ("did something change on their side?").No findings
Why it is in this group
- Nothing in its name or description says what it does, so it is counted as a write.
- The server marks it read-only (readOnlyHint: true), but an annotation never lowers a class.
Description, as the server wrote it Use this when a supplier agent or API you depend on started failing ("did something change on their side?"). Pass its `supplier_url` and the `failure`. Returns the diagnosis plus changes GAIP observed around then (correlated in time, not causation), with receipts. Free, read-only, no account; inputs must be public and non-personal.
Annotations the server declares
- destructiveHint: false
- idempotentHint: false
- openWorldHint: false
- readOnlyHint: true
gaip_tools_changed_sinceUse this before connecting to an MCP server: did its tools change since your `since` date?Can send data outNo findings
Why it is in this group
- Nothing in its name or description says what it does, so it is counted as a write.
- The server marks it read-only (readOnlyHint: true), but an annotation never lowers a class.
What it can reach
- outboundThe input takes a URL (input.url): a request to a URL can carry data to whoever runs it.
Description, as the server wrote it Use this before connecting to an MCP server: did its tools change since your `since` date? Pass its `url`, optionally your tools/list `tools` or a digest. Returns dates, change classes, stored hashes and whether yours match the latest GAIP observed. Facts, not a verdict. Free, read-only, no account; inputs must be public and non-personal.
Annotations the server declares
- destructiveHint: false
- idempotentHint: true
- openWorldHint: false
- readOnlyHint: true
gaip_watch_eventsUse this when asked what changed on a watched agent or service.No findings
Why it is in this group
- Nothing in its name or description says what it does, so it is counted as a write.
- The server marks it read-only (readOnlyHint: true), but an annotation never lowers a class.
Description, as the server wrote it Use this when asked what changed on a watched agent or service. Pass your `continuity_handle` and optionally a `watch_id`. Returns recent change events: verdict before and after, a change summary, the receipt and webhook delivery status. Free, read-only, no account; inputs must be public and non-personal.
Annotations the server declares
- destructiveHint: false
- idempotentHint: true
- openWorldHint: false
- readOnlyHint: true
gaip_watch_registerUse this when asked to keep an eye on an agent or MCP server ("tell me if it changes or goes down").Reads text others writeCan send data outNo findings
Why it is in this group
- Nothing in its name or description says what it does, so it is counted as a write.
What it can reach
- untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.
- outboundThe input takes a URL (input.target_url): a request to a URL can carry data to whoever runs it.
Description, as the server wrote it Use this when asked to keep an eye on an agent or MCP server ("tell me if it changes or goes down"). Pass the `continuity_handle` from any GAIP result, `kind` and `target_url`. GAIP re-checks every 24 hours (3 free watches), receipts each change and can call your webhook. Free, no account; records a GAIP receipt only. Inputs must be public and non-personal.
Annotations the server declares
- destructiveHint: false
- idempotentHint: false
- openWorldHint: true
- readOnlyHint: false
Only reads 2Show 2 read-only tools
gaip_observatory_lookupUse this when asked about an MCP server's or agent's history ("has this server changed?", "was it up last week?").Reads text others writeCan send data outNo findings
Why it is in this group
- The name "gaip_observatory_lookup" says "lookup": a word for reading.
What it can reach
- untrusted-inputThe input takes a URL (input.url), and the tool reads it: it reads web pages, which anyone can write.
- outboundThe input takes a URL (input.url): a request to a URL can carry data to whoever runs it.
Claude Code: ask, because it is tagged outbound.
Description, as the server wrote it Use this when asked about an MCP server's or agent's history ("has this server changed?", "was it up last week?"). Pass its `url` or `name`. Returns dated facts the Agent Observatory observed: status history, manifest hash, conformance codes, change log. Not an assessment. Free, read-only, no account; inputs must be public and non-personal.
Annotations the server declares
- destructiveHint: false
- idempotentHint: true
- openWorldHint: false
- readOnlyHint: true
gaip_watch_listUse this when asked which agents or services you are watching.No findings
Why it is in this group
- The name "gaip_watch_list" says "list": a word for reading.
Description, as the server wrote it Use this when asked which agents or services you are watching. Pass your `continuity_handle`. Returns each watch's kind, target, last verdict, last and next check and webhook endpoint (never the secret). Free, read-only, no account; inputs must be public and non-personal.
Annotations the server declares
- destructiveHint: false
- idempotentHint: true
- openWorldHint: false
- readOnlyHint: true
This is not a safety rating, and no finding is not a safety guarantee. The groups and findings come from fixed checks of the tool definitions the server returned on the date shown, not from its code or what it does when called. How the checks work.
Exfil surface
Which tools read private data, which read text someone else can write, and which can send data out. One of each is a path a prompt injection can use.
30 paths in this server alone from private data, through text someone else wrote, to a way out. Denying 1 private-data tool breaks every one.
- Reads private data1
- Reads files, mail, messages, databases, secrets, calendars or repositories.
- Reads text others write5
- Brings in text someone else can write: web pages, inbound email, messages, issues and comments.
- Can send data out6
- Can carry data to people or systems: it sends, posts or publishes, or takes a recipient, channel or URL.
{
"permissions": {
"deny": [
"mcp__gaip-integration-protocol__gaip_compare_supplier_snapshots"
]
}
}Other servers in the same agent add their own tools to these paths. Audit your whole setup to see them together.
Remote posture
What anyone on the internet sees at the remote endpoint: an anonymous initialize and tools/list, and the public OAuth metadata.
Looked at Nov 20, 2026, 19:23 UTC, unchanged since Oct 22, 2026. No credentials were sent and no tool was called. What is looked at.
- Reachable
- yes
- HTTP status of an anonymous initialize
- 200
- tools/list served without credentials
- yes
- Tools that change things, listed without credentials
- 11
- Protocol version
- 2025-06-18
- WWW-Authenticate on 401
- not observed
- resource_metadata in WWW-Authenticate
- not observed
- RFC 9728 protected resource metadata
- not observed
- Authorization servers
- not observed
- Legacy HTTP+SSE transport
- no
- Dynamic client registration
- not observed
- Client ID metadata documents
- not observed
- HighThe whole server
The endpoint served tools/list without credentials, and the list has 11 tools that change things (1 destructive, 1 irreversible, 9 write). Whether a call also works without credentials was not tested: tools are never called.
What this check looks forMatched text
https://www.gaipagents.com/mcp/agents/integration-protocol: tools/list returned 13 tools to an anonymous client, including gaip_watch_remove (destructive), gaip_observatory_capture (irreversible), check_agent_conformance (write), gaip_watch_register (write), gaip_watch_events (write) and 6 more
Findings
Matches of fixed checks, with the text they matched.
The server marks "gaip_compare_supplier_snapshots" as not read-only (readOnlyHint: false), and its words say it reads.
What this check looks forMatched text
annotations: {"readOnlyHint":false,"destructiveHint":false,"idempotentHint":false,"openWorldHint":false}; The name "gaip_compare_supplier_snapshots" says "compare": a word for reading.
History
Changes to its tool definitions and to its remote endpoint, newest first.
No change seen yet. Its tools are rescanned every day, and a remote endpoint is looked at once a week.
Install
Pinned to version 1.5.0. Values in angle brackets are yours to fill in.
claude mcp add --transport http gaip-integration-protocol https://www.gaipagents.com/mcp/agents/integration-protocolclaude mcp add --transport http gaip-integration-protocol https://www.gaipagents.com/mcp/agents/integration-protocolRun in a terminal.
Claude Code permissions
Allow the reads, ask before a change or a way out, deny the deletes. Paste into .claude/settings.json.
{
"permissions": {
"allow": [
"mcp__gaip-integration-protocol__gaip_watch_list"
],
"ask": [
"mcp__gaip-integration-protocol__check_agent_conformance",
"mcp__gaip-integration-protocol__gaip_watch_register",
"mcp__gaip-integration-protocol__gaip_watch_events",
"mcp__gaip-integration-protocol__gaip_observatory_lookup",
"mcp__gaip-integration-protocol__gaip_manifest_changes",
"mcp__gaip-integration-protocol__gaip_observatory_capture",
"mcp__gaip-integration-protocol__gaip_tools_changed_since",
"mcp__gaip-integration-protocol__gaip_check_permission",
"mcp__gaip-integration-protocol__gaip_compare_supplier_snapshots",
"mcp__gaip-integration-protocol__gaip_diagnose_integration_failure",
"mcp__gaip-integration-protocol__gaip_supplier_reliability"
],
"deny": [
"mcp__gaip-integration-protocol__gaip_watch_remove"
]
}
}Tools that only read are allowed, tools that change things ask first, tools that delete or cancel are denied. A read that can send data out (it takes a URL, say) asks first too. It assumes the server is installed as gaip-integration-protocol. Review it before you use it.
Capability tags behind 7 rules
- mcp__gaip-integration-protocol__check_agent_conformance untrusted-input, outbound
- mcp__gaip-integration-protocol__gaip_watch_register untrusted-input, outbound
- mcp__gaip-integration-protocol__gaip_observatory_lookup untrusted-input, outbound
- mcp__gaip-integration-protocol__gaip_observatory_capture untrusted-input, outbound
- mcp__gaip-integration-protocol__gaip_tools_changed_since outbound
- mcp__gaip-integration-protocol__gaip_check_permission untrusted-input, outbound
- mcp__gaip-integration-protocol__gaip_compare_supplier_snapshots private-data
Pin
Keep the decision: fail CI when the tools change, follow every change, or show a badge.
Pin in CI
Fail a build when the tools you reviewed are not the tools the server serves today.
curl -s "https://radar.anlyon.com/api/v1/servers/io.github.tjcgraham-rgb/gaip-integration-protocol/verify?hash=e7a554beaa49fa5ca43050a88cf06a4605e4f89c50945bec07a3e27db18f6387" | jq -e '.changed == false'Pinned hash e7a554beaa49, from Nov 21, 2026, 08:18 UTC.
Watch
Get told when its tools change: a new tool, a new description, a new class.
https://radar.anlyon.com/servers/io.github.tjcgraham-rgb/gaip-integration-protocol/changes.xmlAn Atom feed. Works in any feed reader or Slack's RSS app.
Badge
[](https://radar.anlyon.com/servers/io.github.tjcgraham-rgb/gaip-integration-protocol)Source: MCP Registry · repository. Synced Oct 21, 2026, 18:50 UTC.
GAIP Integration & Protocol Agent MCP server: questions and answers
Answered from the scan above, not from the publisher's description.
Is the GAIP Integration & Protocol Agent MCP server safe?
- No scan can promise that. What its definitions show: 13 tools, 11 of which can change, send or delete something, and 1 finding from fixed checks. Allow the reads, ask before each change, and deny the deletes; the permissions block on this page does that.
Can the GAIP Integration & Protocol Agent MCP server delete data?
- Yes. 1 tool can delete, cancel or revoke something: gaip_watch_remove.
Can the GAIP Integration & Protocol Agent MCP server do things that cannot be undone?
- Yes. 1 tool can send, pay, publish or merge: gaip_observatory_capture.
What tools does the GAIP Integration & Protocol Agent MCP server have?
- 13 tools: 1 destructive, 1 irreversible, 9 write, 2 read. gaip_watch_remove, gaip_observatory_capture, check_agent_conformance, gaip_watch_register, gaip_watch_events, gaip_manifest_changes, gaip_tools_changed_since, gaip_check_permission and 5 more.
Does the GAIP Integration & Protocol Agent MCP server need sign-in or OAuth?
- Its remote endpoint listed its tools to an anonymous client, without sign-in.
Does the GAIP Integration & Protocol Agent MCP server send data to outside services?
- 6 tools can carry data out (it sends, posts or publishes, or takes a recipient, channel or URL): check_agent_conformance, gaip_watch_register, gaip_observatory_lookup, gaip_observatory_capture, gaip_tools_changed_since and gaip_check_permission.
Has the GAIP Integration & Protocol Agent MCP server changed its tools?
- Not since Anlyon Radar started tracking it. Last scanned Nov 21, 2026.