AgentBreakWatch

io.github.joeyg6393/agentbreakwatch-mcpjoeyg6393streamable-httpv1.1.0Scanned 1 hour ago

Audit AI Agent SDKs and MCP servers for breaking changes, schema drift, and migration fixes.

4 tools: 1 change things and 3 only read.

Reads private data · 0
Reads text others write · 0
Can send data out · 0
2 findings, 2 high

Anlyon alternative

1 of these tools change something. Run those as governed actions.

Declare what an action changes, and Anlyon puts it under a shared limit, binds every approval to the exact request, and grades the outcome. It works on any HTTPS API you can describe as an action.

Anlyon does not run, proxy or govern this MCP server. A tool your own code calls directly is not routed through Anlyon and is not governed by it.

Import an OpenAPI spec →

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.

4 tools

Changes things 1Creates or changes something that can be changed back.Claude Code: ask

check_stack_driftAudit a package.json, requirements.txt, or .mcp.json configuration string against 256 live AI Agent SDKs and MCP servers.2 findings
Why it is in this group
- Nothing in its name or description says what it does, so it is counted as a write.
Findings
- High
  The description refers to ".mcp.json".
  Matched text
  description: "...it a package.json, requirements.txt, or .mcp.json configuration string against 256 live A..."
  What this check looks for
- High
  The input schema at properties.manifest.description refers to ".mcp.json".
  Matched text
  inputSchema.properties.manifest.description: "Raw contents of .mcp.json, package.json, requirements.txt, or a n..."
  What this check looks for
Description, as the server wrote it
Audit a package.json, requirements.txt, or .mcp.json configuration string against 256 live AI Agent SDKs and MCP servers. Flags breaking major version deltas, unpinned @latest MCP servers, and outdated dependencies with exact pin commands.

Only reads 3Looks things up and changes nothing.Show 3 read-only tools

get_archived_mcp_replacementsGet the migration matrix for archived or frozen @modelcontextprotocol/server-* reference packages and their official 2026 vendor-maintained replacements.No findings
Why it is in this group
- The name "get_archived_mcp_replacements" says "get": a word for reading.
Description, as the server wrote it
Get the migration matrix for archived or frozen @modelcontextprotocol/server-* reference packages and their official 2026 vendor-maintained replacements.
get_package_stabilityLook up live npm or PyPI version, publish timestamp, 90-day release velocity, stability score (0-100), MCP tools/list schema drift risk, and before/after migration code for a speci…No findings
Why it is in this group
- The name "get_package_stability" says "get": a word for reading.
Description, as the server wrote it
Look up live npm or PyPI version, publish timestamp, 90-day release velocity, stability score (0-100), MCP tools/list schema drift risk, and before/after migration code for a specific AI Agent SDK or MCP server.
list_volatile_packagesList the fastest-changing or highest schema-drift-risk AI Agent SDKs and MCP servers from the live AgentBreakWatch telemetry index.No findings
Why it is in this group
- The name "list_volatile_packages" says "list": a word for reading.
Description, as the server wrote it
List the fastest-changing or highest schema-drift-risk AI Agent SDKs and MCP servers from the live AgentBreakWatch telemetry index.

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.

No private-data, untrusted-input, outbound path in this server alone: no tool is tagged private-data or untrusted-input or outbound.

Reads private data0: Reads files, mail, messages, databases, secrets, calendars or repositories.
Reads text others write0: Brings in text someone else can write: web pages, inbound email, messages, issues and comments.
Can send data out0: Can carry data to people or systems: it sends, posts or publishes, or takes a recipient, channel or URL.

Other servers in the same agent add their own tools to these paths. Audit your whole setup to see them together.

Findings

Matches of fixed checks, with the text they matched.

2 high

Highcheck_stack_drift
The description refers to ".mcp.json".
Matched text
description: "...it a package.json, requirements.txt, or .mcp.json configuration string against 256 live A..."
What this check looks for
Highcheck_stack_drift
The input schema at properties.manifest.description refers to ".mcp.json".
Matched text
inputSchema.properties.manifest.description: "Raw contents of .mcp.json, package.json, requirements.txt, or a n..."
What this check looks for

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.

All changes to this server

Install

Pinned to version 1.1.0. Values in angle brackets are yours to fill in.

claude mcp add --transport http agentbreakwatch https://agentbreakwatch.com/api/mcp

claude mcp add --transport http agentbreakwatch https://agentbreakwatch.com/api/mcp

Run 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.

.claude/settings.json

{
  "permissions": {
    "allow": [
      "mcp__agentbreakwatch__get_package_stability",
      "mcp__agentbreakwatch__list_volatile_packages",
      "mcp__agentbreakwatch__get_archived_mcp_replacements"
    ],
    "ask": [
      "mcp__agentbreakwatch__check_stack_drift"
    ],
    "deny": []
  }
}

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 agentbreakwatch. Review it before you use it.

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://anlyon.dev/api/v1/servers/io.github.joeyg6393/agentbreakwatch-mcp/verify?hash=93d53aeb9758e84ea4edef87f1e0a43c52e856b9e5c0bf1e7283ca650566b90b" | jq -e '.changed == false'

Pinned hash 93d53aeb9758, from Oct 25, 2026, 13:18 UTC.

Watch

Get told when its tools change: a new tool, a new description, a new class.

https://anlyon.dev/servers/io.github.joeyg6393/agentbreakwatch-mcp/changes.xml

An Atom feed. Works in any feed reader or Slack's RSS app.

Badge

[![anlyon.dev](https://anlyon.dev/badge/io.github.joeyg6393/agentbreakwatch-mcp.svg)](https://anlyon.dev/servers/io.github.joeyg6393/agentbreakwatch-mcp)

Source: MCP Registry · repository. Synced Oct 25, 2026, 13:17 UTC.