Skip to content

Agent Router MCP server

io.github.Tim-Projekt/agent-routerTim-Projektstreamable-httpv1.0.0Scanned

Discover and call AI agents via MCP. Supports A2A agents and platform agents with async tasks.

27 tools: 21 change things and 6 only read.

Tools: 0 delete, 0 irreversible, 21 write, 6 read.
  • Reads private data · 0
  • Reads text others write · 0
  • Can send data out · 0
  • No findings

Tools listed without sign-in. Remote posture

With Anlyon

21 tools here change things. Govern them.

  • Anlyon lists the tools and adopts none.
  • Each tool you adopt gets approvals and receipts.
  • Your agent never holds the key.
Govern this server →

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 Agent Router MCP server can do

The Agent Router MCP server is published by the GitHub account Tim-Projekt. It has 27 tools. It can create or change data (21 tools: a2a_call_agent, sandboxcodingagent, softwareengineeringexpert, testagent and 17 more), but cannot delete data. 6 tools only read.

None of its tools reads private data, brings in text other people write, or sends data out.

None of the fixed checks matched anything in its tool definitions. That is not a safety guarantee.

Its remote endpoint lists its tools without sign-in, and its tools changed once since Anlyon Radar first saw them.

Last scanned Nov 16, 2026, version 1.0.0.

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.

27 tools

Changes things 21Claude Code: ask
  • a2a_call_agentCall any public A2A-compatible agent directly by its endpoint URL.No findings

    Why it is in this group

    • The name "a2a_call_agent" says "call": a word for a change.
    Description, as the server wrote it
    Call any public A2A-compatible agent directly by its endpoint URL. Returns the agent's response immediately (synchronous — no task_id needed). Get agent_url values from discover_agents (look for ENDPOINT in results). Works with any agent on a2aregistry.org or any A2A JSON-RPC endpoint. Free, no login required. Args: agent_url: The A2A endpoint URL (from discover_agents ENDPOINT field). message: The message or task to send to the agent. context_id: Optional — pass the context_id from a prior response to continue a multi-turn conversation. timeout_seconds: Seconds to wait for a response (default 30, max 120).
  • browsernavigationagentPerforms goal-oriented web navigation to complete tasks using a browser tool.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
    Performs goal-oriented web navigation to complete tasks using a browser tool. Accesses websites, searches, clicks links, interacts with forms, and extracts relevant content. Produces structured results based solely on navigated information and supports the Main Agent with verified web-based insights. Expected Runtime: ~30s.
  • campbuddyPlan complete camping trips based on destination, weather, experience level, group size, budget, and activities.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
    Plan complete camping trips based on destination, weather, experience level, group size, budget, and activities. Generate packing lists, meal plans, safety checklists, itineraries, campsite recommendations, and emergency preparation. Expected Runtime: ~30s.
  • exploitagentAgent created via direct DB write Expected Runtime: ~30s.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
    Agent created via direct DB write Expected Runtime: ~30s.
  • financial_agentfull financial agent Expected Runtime: ~60s.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
    full financial agent Expected Runtime: ~60s.
  • firstprinciplesanalystDeconstructs complex tasks or problems using a first-principles approach.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
    Deconstructs complex tasks or problems using a first-principles approach. Identifies fundamental components, dependencies, and leverage points, and generates atomic-level insights and strategic perspectives. Provides guidance and clarity without delivering ready-made solutions, supporting the Main Agent in understanding and structuring the problem. Expected Runtime: ~30s.
  • hpo_demo_echo_agentDemo agent that echoes back any text query it receives.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
    Demo agent that echoes back any text query it receives. Useful for testing A2A webhook integrations. Expected Runtime: ~10s.
  • luraph_deobfuscatorcodes Expected Runtime: ~30s.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
    codes Expected Runtime: ~30s.
  • nexus_strategy_analystAutonomous strategic analysis framed in NEXUS's lens: competitive positioning, portfolio economics, and defensibility.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
    Autonomous strategic analysis framed in NEXUS's lens: competitive positioning, portfolio economics, and defensibility. Returns structured, actionable strategy recommendations for business problems. Expected Runtime: ~20s.
  • reasoningdelegationhighThe high-effort reasoning agent is built for complex tasks that require deep, multi-step analysis and robust problem-solving.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
    The high-effort reasoning agent is built for complex tasks that require deep, multi-step analysis and robust problem-solving. It thoroughly evaluates alternatives, connects multiple sources of information, and carefully reasons through uncertainty before producing an answer. This agent is suited for challenging planning, architecture design, and analytical tasks where accuracy and depth are critical. Expected Runtime: ~60s.
  • reasoningdelegationlowA low-effort reasoning agent designed to handle simple to moderate tasks by performing lightweight reflection and structured thinking.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
    A low-effort reasoning agent designed to handle simple to moderate tasks by performing lightweight reflection and structured thinking. It provides quick, cost-efficient reasoning support when full deep analysis is not required. The agent can be called upon to clarify problems, outline steps, and make small to mid-level decisions with minimal computational overhead. Expected Runtime: ~45s.
  • reasoningdelegationmediumThe medium-effort reasoning agent is designed to handle moderate to complex tasks by applying structured, multi-step thinking and deeper analysis.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
    The medium-effort reasoning agent is designed to handle moderate to complex tasks by applying structured, multi-step thinking and deeper analysis. It balances efficiency and depth, enabling reliable problem decomposition, evaluation of alternatives, and coherent decision-making. This agent is suitable for tasks that require thoughtful reasoning but do not justify maximum computational intensity. Expected Runtime: ~50s.
  • researchagentConducts thorough, iterative internet research on a given topic.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
    Conducts thorough, iterative internet research on a given topic. Identifies key terms, subtopics, facts, studies, and current developments. Compares sources, validates information, and highlights uncertainties. Produces structured, high-quality insights and open questions to support the Main Agent’s understanding. Expected Runtime: ~60s.
  • sandboxcodingagentExecutes Python or JavaScript code in an isolated, stateless sandbox to validate, test, or solve tasks.No findings

    Why it is in this group

    • The description's first verb is "execute": a word for a change.
    Description, as the server wrote it
    Executes Python or JavaScript code in an isolated, stateless sandbox to validate, test, or solve tasks. Produces concrete, structured outputs from actual code execution, systematically tests logic and clearly identifies what is achievable within sandbox constraints. Supports the Main Agent with verified results and actionable insights. Expected Runtime: ~45s.
  • silpo_home_restaurantCreates a complete home restaurant experience from the household fridge, guest count, cuisine, time and budget.No findings

    Why it is in this group

    • The description's first verb is "create": a word for a change.
    Description, as the server wrote it
    Creates a complete home restaurant experience from the household fridge, guest count, cuisine, time and budget. Builds menus, identifies missing ingredients, prepares a transparent Silpo shopping handoff, generates a guest table passport with ingredients and allergens, and drafts a message for friends. Sponsored products are labelled and only shown when relevant. Expected Runtime: ~15s.
  • smart_fridge___nutritionTracks household food inventory, expiry risks and replenishment needs; considers user food restrictions and preferences; prepares transparent product recommendations, a draft Silpo…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
    Tracks household food inventory, expiry risks and replenishment needs; considers user food restrictions and preferences; prepares transparent product recommendations, a draft Silpo basket and delivery options. Sponsored products are always labelled and shown only when genuinely suitable. Cart changes require explicit user approval. Expected Runtime: ~15s.
  • softwareengineeringexpertPerforms expert-level software engineering analysis to solve tasks using repository and code evidence as the primary source.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
    Performs expert-level software engineering analysis to solve tasks using repository and code evidence as the primary source. Investigates architecture, dependencies, design patterns, and code quality to generate structured technical insights, trade-offs, and strategic recommendations. Supports the Main Agent with actionable engineering guidance without speculation. Can't execute code. Expected Runtime: ~50s.
  • stock_marketFundamental Analysis Expected Runtime: ~30s.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
    Fundamental Analysis Expected Runtime: ~30s.
  • taskplannerGenerates structured, iterative execution plans from complex user requests or problems.No findings

    Why it is in this group

    • The description's first verb is "generate": a word for a change.
    Description, as the server wrote it
    Generates structured, iterative execution plans from complex user requests or problems. Suggests task breakdowns, phase sequencing, and knowledge flow, but does not independently execute or delegate tasks to other agents. Supports the Main Agent in preparing actionable plans. Expected Runtime: ~30s.
  • testagentDelegates a task to the specialized agent: TestAgent Expected Runtime: ~30s.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
    Delegates a task to the specialized agent: TestAgent Expected Runtime: ~30s.
Only reads 6Show 6 read-only tools
  • constructivecriticAnalyzes complex approaches and strategies by identifying strengths and weaknesses.No findings

    Why it is in this group

    • The description's first verb is "analyze": a word for reading.
    Description, as the server wrote it
    Analyzes complex approaches and strategies by identifying strengths and weaknesses. Provides structured feedback with concrete suggestions for improvement and potential alternatives. Expected Runtime: ~30s.
  • discover_agentsFind agents to call — both platform agents and public A2A registry agents.No findings

    Why it is in this group

    • The description's first verb is "find": a word for reading.
    Description, as the server wrote it
    Find agents to call — both platform agents and public A2A registry agents. Returns two types: • TYPE=platform — built-in agents, call via their MCP tool name (async, returns task_id → use wait_for_task) • TYPE=a2a_registry — public agents from a2aregistry.org, call via a2a_call_agent(agent_url=ENDPOINT, message='...') (sync, returns immediately) Registry agents are filtered by the registry's own is_healthy flag. Each result shows UPTIME and LATENCY from the registry's own reported metrics. Free. Args: query: Keywords to filter by capability (e.g. 'weather', 'web scraping', 'research'). Leave empty to browse top agents. limit: Max results to return (default 10, max 25).
  • get_skillFetch a ClawHub skill by slug — metadata plus optional SKILL.md body.No findings

    Why it is in this group

    • The name "get_skill" says "get": a word for reading.
    Description, as the server wrote it
    Fetch a ClawHub skill by slug — metadata plus optional SKILL.md body. Use the slug from search_skills results (e.g. 'agent-browser-clawdbot'). The response includes version, stats, install command, and the SKILL.md content agents can follow to adopt the skill's behavior. Free to use — no API key required. Args: slug: ClawHub skill slug (lowercase, from search_skills). include_content: If true (default), include SKILL.md text. Set false for metadata only.
  • scientificresearchagentAnalyzes tasks, questions, or problems strictly from a scientific perspective.No findings

    Why it is in this group

    • The description's first verb is "analyze": a word for reading.
    Description, as the server wrote it
    Analyzes tasks, questions, or problems strictly from a scientific perspective. Retrieves, synthesizes, and evaluates evidence from academic literature across relevant domains. Produces structured, expert-level insights while distinguishing established findings, emerging evidence, and uncertainties. Supports the Main Agent with scientifically rigorous reasoning and recommendations. Expected Runtime: ~60s.
  • search_skillsSemantically search the live ClawHub skills registry (clawhub.ai).No findings

    Why it is in this group

    • The name "search_skills" says "search": a word for reading.
    Description, as the server wrote it
    Semantically search the live ClawHub skills registry (clawhub.ai). ClawHub hosts OpenClaw SKILL.md packages — behavior templates, tool integrations, and agent capabilities. Results include slug, author, summary, install hint, and canonical URL. Follow up with get_skill(slug) to read the full SKILL.md. Free to use — no API key required. Args: query: Natural language description of the capability you need (e.g. 'browser automation', 'send emails via gmail'). top_k: Number of results to return (default: 5, max: 10).
  • wait_for_taskWait for a platform agent task to complete and return its result.No findings

    Why it is in this group

    • The name "wait_for_task" says "wait": a word for reading.
    Description, as the server wrote it
    Wait for a platform agent task to complete and return its result. Only needed when a platform agent tool returned STATUS=RUNNING with a task_id (i.e. the task was still running after the initial 50s inline wait). NOT needed when the tool already returned STATUS=COMPLETED or STATUS=FAILED. NOT needed for a2a_call_agent — that always returns directly. Args: task_id: The task UUID from a platform agent response with STATUS=RUNNING. max_wait_seconds: Max seconds to wait (default 45, max 300).

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.

Remote posture

What anyone on the internet sees at the remote endpoint: an anonymous initialize and tools/list, and the public OAuth metadata.

Tools listed without sign-in

Looked at Nov 20, 2026, 19:00 UTC, unchanged since Nov 6, 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
21
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
  • WarnThe whole server

    The endpoint served tools/list without credentials, and the list has 21 tools that change things (21 write). Whether a call also works without credentials was not tested: tools are never called.

    Matched text

    https://agent-router.org/mcp-http/mcp: tools/list returned 27 tools to an anonymous client, including a2a_call_agent (write), sandboxcodingagent (write), softwareengineeringexpert (write), testagent (write), exploitagent (write) and 16 more

    What this check looks for

Posture changes

  1. Nov 6, 2026
    • Tools that change things, listed without credentials: 20 to 21

Findings

Matches of fixed checks, with the text they matched.

No check matched anything in these tool definitions. That is not a safety guarantee.

History

Changes to its tool definitions and to its remote endpoint, newest first.

  1. Tools1 added.+1 addedView diff
  2. PostureRemote posture changed. Tools that change things, listed without credentials: 20 -> 21.View diff
    • Tools that change things, listed without credentials: 20 to 21

Install

Pinned to version 1.0.0. Values in angle brackets are yours to fill in.

claude mcp add --transport http agent-router https://agent-router.org/mcp-http/mcp
claude mcp add --transport http agent-router https://agent-router.org/mcp-http/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__agent-router__wait_for_task",
      "mcp__agent-router__discover_agents",
      "mcp__agent-router__search_skills",
      "mcp__agent-router__get_skill",
      "mcp__agent-router__scientificresearchagent",
      "mcp__agent-router__constructivecritic"
    ],
    "ask": [
      "mcp__agent-router__a2a_call_agent",
      "mcp__agent-router__sandboxcodingagent",
      "mcp__agent-router__softwareengineeringexpert",
      "mcp__agent-router__testagent",
      "mcp__agent-router__exploitagent",
      "mcp__agent-router__taskplanner",
      "mcp__agent-router__researchagent",
      "mcp__agent-router__browsernavigationagent",
      "mcp__agent-router__reasoningdelegationlow",
      "mcp__agent-router__firstprinciplesanalyst",
      "mcp__agent-router__reasoningdelegationmedium",
      "mcp__agent-router__reasoningdelegationhigh",
      "mcp__agent-router__campbuddy",
      "mcp__agent-router__nexus_strategy_analyst",
      "mcp__agent-router__financial_agent",
      "mcp__agent-router__silpo_home_restaurant",
      "mcp__agent-router__hpo_demo_echo_agent",
      "mcp__agent-router__russian_website_legal_auditor",
      "mcp__agent-router__smart_fridge___nutrition",
      "mcp__agent-router__stock_market",
      "mcp__agent-router__luraph_deobfuscator"
    ],
    "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 agent-router. Review it before you use it.

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.Tim-Projekt/agent-router/verify?hash=ed04d32379b0667ddc465133b179a3051b8736067ef83ebaa478b1e9b7fb0cb0" | jq -e '.changed == false'

Pinned hash ed04d32379b0, from Nov 16, 2026, 06:18 UTC.

Watch

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

https://radar.anlyon.com/servers/io.github.Tim-Projekt/agent-router/changes.xml

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

Badge

Anlyon Radar badge for io.github.Tim-Projekt/agent-router
[![Anlyon Radar](https://radar.anlyon.com/badge/io.github.Tim-Projekt/agent-router.svg)](https://radar.anlyon.com/servers/io.github.Tim-Projekt/agent-router)

Source: MCP Registry · repository. Synced Oct 21, 2026, 18:21 UTC.

Agent Router MCP server: questions and answers

Answered from the scan above, not from the publisher's description.

Is the Agent Router MCP server safe?

No scan can promise that. What its definitions show: 27 tools, 21 of which can change, send or delete something, and 0 findings from fixed checks. Allow the reads, ask before each change, and deny the deletes; the permissions block on this page does that.

Can the Agent Router MCP server delete data?

No tool is classed as destructive: going by their definitions, none of them deletes, cancels or revokes anything.

Can the Agent Router MCP server do things that cannot be undone?

No tool is classed as irreversible: none of them sends, pays, publishes or merges, going by their definitions.

What tools does the Agent Router MCP server have?

27 tools: 21 write, 6 read. a2a_call_agent, sandboxcodingagent, softwareengineeringexpert, testagent, exploitagent, taskplanner, researchagent, browsernavigationagent and 19 more.

Does the Agent Router MCP server need sign-in or OAuth?

Its remote endpoint listed its tools to an anonymous client, without sign-in.

Does the Agent Router MCP server send data to outside services?

None of its tools sends, posts or publishes, or takes a recipient, channel or URL.

Has the Agent Router MCP server changed its tools?

Yes, once since Anlyon Radar started tracking it, most recently on Nov 16, 2026. Every change is on this page and in its Atom feed.