Skip to content

Roundtable MCP server

now.roundtable.mcp/roundtablemcp.roundtable.nowstreamable-httpv1.0.2Scanned

Multi-model AI debates: GPT-4o, Claude, Gemini & 200+ models discuss, then synthesize insight.

13 tools: 8 change things and 5 only read.

Tools: 0 delete, 0 irreversible, 8 write, 5 read.
  • Reads private data · 1
  • Reads text others write · 7
  • Can send data out · 6
  • No findings

Tools listed without sign-in. Remote posture

With Anlyon

8 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 Roundtable MCP server can do

The Roundtable MCP server is published by mcp.roundtable.now, a domain the MCP Registry verified. It has 13 tools. It can create or change data (8 tools: check-usage, set-thread-visibility, consult-council, design-architecture and 4 more), but cannot delete data. 5 tools only read.

1 tool reads private data (get-thread-link); 7 tools bring in text other people can write (get-thread-link, consult-council, design-architecture and 4 more); 6 tools can send data out of your system (consult-council, design-architecture, review-code 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.

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.

Last scanned Nov 18, 2026, version 1.0.2.

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

Changes things 8Claude Code: ask
  • assess-tradeoffsTradeoff assessment council.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.webhook_url): a request to a URL can carry data to whoever runs it.
    Description, as the server wrote it
    Tradeoff assessment council. Pragmatist, Skeptic, and Futurist evaluate options from different angles — short-term vs long-term, risk vs reward, simplicity vs flexibility. Output as pros-cons.

    Annotations the server declares

    • destructiveHint: false
    • idempotentHint: false
    • openWorldHint: true
    • readOnlyHint: false
  • check-usageCheck your remaining credits, usage limits, and plan infoNo 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
    Check your remaining credits, usage limits, and plan info

    Annotations the server declares

    • destructiveHint: false
    • idempotentHint: true
    • openWorldHint: false
    • readOnlyHint: true
  • consult-councilConsult the AI coding council — multiple models discuss your engineering question sequentially (each sees prior responses), then a moderator synthesizes.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.webhook_url): a request to a URL can carry data to whoever runs it.
    Description, as the server wrote it
    Consult the AI coding council — multiple models discuss your engineering question sequentially (each sees prior responses), then a moderator synthesizes. Auto-mode by default — AI picks optimal models, roles, and conversation mode from your prompt. Provide explicit models to override (manual mode). Fully configurable: mode, format, roles, models, thinking level.

    Annotations the server declares

    • destructiveHint: false
    • idempotentHint: false
    • openWorldHint: true
    • readOnlyHint: false
  • debug-issueDebugging council.Reads text others writeCan send data outNo findings

    Why it is in this group

    • The input takes a statement or command to run (input.code), which can change anything it can reach.

    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.webhook_url): a request to a URL can carry data to whoever runs it.
    Description, as the server wrote it
    Debugging council. Root Cause Analyst, Systems Engineer, and Edge Case Investigator collaboratively diagnose bugs, analyze errors, and propose fixes.

    Annotations the server declares

    • destructiveHint: false
    • idempotentHint: false
    • openWorldHint: true
    • readOnlyHint: false
  • design-architectureArchitecture design council.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.webhook_url): a request to a URL can carry data to whoever runs it.
    Description, as the server wrote it
    Architecture design council. Systems Architect, Infrastructure Engineer, and DX Advocate evaluate your system design. Always uses high thinking for maximum depth. Output as ADR.

    Annotations the server declares

    • destructiveHint: false
    • idempotentHint: false
    • openWorldHint: true
    • readOnlyHint: false
  • plan-implementationImplementation planning council.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.webhook_url): a request to a URL can carry data to whoever runs it.
    Description, as the server wrote it
    Implementation planning council. Tech Lead, Senior Engineer, and QA Strategist break down a feature into actionable steps, identify risks, and define acceptance criteria. Output as ADR.

    Annotations the server declares

    • destructiveHint: false
    • idempotentHint: false
    • openWorldHint: true
    • readOnlyHint: false
  • review-codeCode review council.Reads text others writeCan send data outNo findings

    Why it is in this group

    • The input takes a statement or command to run (input.code), which can change anything it can reach.

    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.webhook_url): a request to a URL can carry data to whoever runs it.
    Description, as the server wrote it
    Code review council. Senior Engineer, Security Reviewer, and Performance Analyst analyze your code and a moderator synthesizes their findings.

    Annotations the server declares

    • destructiveHint: false
    • idempotentHint: false
    • openWorldHint: true
    • readOnlyHint: false
  • set-thread-visibilitySet a thread as public or private.No findings

    Why it is in this group

    • The name "set-thread-visibility" says "set": a word for a change.
    Description, as the server wrote it
    Set a thread as public or private. Public threads can be shared via URL. Use session_id from a previous debate session.

    Annotations the server declares

    • destructiveHint: false
    • idempotentHint: true
    • openWorldHint: false
    • readOnlyHint: false
Only reads 5Show 5 read-only tools
  • get-logsQuery structured logs from your MCP tool executions.No findings

    Why it is in this group

    • The name "get-logs" says "get": a word for reading.
    Description, as the server wrote it
    Query structured logs from your MCP tool executions. Filter by session, severity level, event type, and time range. Useful for debugging and monitoring tool usage.

    Annotations the server declares

    • destructiveHint: false
    • idempotentHint: true
    • openWorldHint: false
    • readOnlyHint: true
  • get-sessionGet full details of a previous MCP session by ID.No findings

    Why it is in this group

    • The name "get-session" says "get": a word for reading.
    Description, as the server wrote it
    Get full details of a previous MCP session by ID. Returns the complete result including participant responses and moderator synthesis. Use list-sessions first to find session IDs.

    Annotations the server declares

    • destructiveHint: false
    • idempotentHint: true
    • openWorldHint: false
    • readOnlyHint: true
  • list-modelsList available AI models grouped by thinking level (low/medium/high).No findings

    Why it is in this group

    • The name "list-models" says "list": a word for reading.
    Description, as the server wrote it
    List available AI models grouped by thinking level (low/medium/high). Shows default models, credit costs, capabilities for each tier. Use this before consult to understand model options.

    Annotations the server declares

    • destructiveHint: false
    • idempotentHint: true
    • openWorldHint: false
    • readOnlyHint: true
  • list-sessionsList your previous MCP tool sessions.No findings

    Why it is in this group

    • The name "list-sessions" says "list": a word for reading.
    Description, as the server wrote it
    List your previous MCP tool sessions. Returns session metadata including prompt, tool used, quality score, and credits consumed. Useful for reviewing past council discussions.

    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.

42 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 write7
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.
deny to break every path
{
  "permissions": {
    "deny": [
      "mcp__roundtable__get-thread-link"
    ]
  }
}

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, 19:16 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
8
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 8 tools that change things (8 write). Whether a call also works without credentials was not tested: tools are never called.

    Matched text

    https://mcp.roundtable.now/mcp: tools/list returned 13 tools to an anonymous client, including check-usage (write), set-thread-visibility (write), consult-council (write), design-architecture (write), review-code (write) and 3 more

    What this check looks for

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.

No change seen yet. Its tools are rescanned every day, and a remote endpoint is looked at once a week.

Install

Pinned to version 1.0.2. Values in angle brackets are yours to fill in.

claude mcp add --transport http roundtable https://mcp.roundtable.now/mcp --header 'X-API-Key: <your-x-api-key>'
claude mcp add --transport http roundtable https://mcp.roundtable.now/mcp --header 'X-API-Key: <your-x-api-key>'

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__roundtable__list-models",
      "mcp__roundtable__list-sessions",
      "mcp__roundtable__get-session",
      "mcp__roundtable__get-logs",
      "mcp__roundtable__get-thread-link"
    ],
    "ask": [
      "mcp__roundtable__check-usage",
      "mcp__roundtable__set-thread-visibility",
      "mcp__roundtable__consult-council",
      "mcp__roundtable__design-architecture",
      "mcp__roundtable__review-code",
      "mcp__roundtable__plan-implementation",
      "mcp__roundtable__debug-issue",
      "mcp__roundtable__assess-tradeoffs"
    ],
    "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 roundtable. Review it before you use it.

Capability tags behind 7 rules
  • mcp__roundtable__get-thread-link private-data, untrusted-input
  • mcp__roundtable__consult-council untrusted-input, outbound
  • mcp__roundtable__design-architecture untrusted-input, outbound
  • mcp__roundtable__review-code untrusted-input, outbound
  • mcp__roundtable__plan-implementation untrusted-input, outbound
  • mcp__roundtable__debug-issue untrusted-input, outbound
  • mcp__roundtable__assess-tradeoffs untrusted-input, outbound

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/now.roundtable.mcp/roundtable/verify?hash=4486a1567d3a0bceb4a1bd0a83e5c9eb48b07fc4104424702d9f7f01ccae3c64" | jq -e '.changed == false'

Pinned hash 4486a1567d3a, from Nov 18, 2026, 02:17 UTC.

Watch

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

https://radar.anlyon.com/servers/now.roundtable.mcp/roundtable/changes.xml

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

Badge

Anlyon Radar badge for now.roundtable.mcp/roundtable
[![Anlyon Radar](https://radar.anlyon.com/badge/now.roundtable.mcp/roundtable.svg)](https://radar.anlyon.com/servers/now.roundtable.mcp/roundtable)

Source: MCP Registry · repository. Synced Oct 21, 2026, 18:54 UTC.

Roundtable MCP server: questions and answers

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

Is the Roundtable MCP server safe?

No scan can promise that. What its definitions show: 13 tools, 8 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 Roundtable MCP server delete data?

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

Can the Roundtable 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 Roundtable MCP server have?

13 tools: 8 write, 5 read. check-usage, set-thread-visibility, consult-council, design-architecture, review-code, plan-implementation, debug-issue, assess-tradeoffs and 5 more.

Does the Roundtable MCP server need sign-in or OAuth?

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

Does the Roundtable 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): consult-council, design-architecture, review-code, plan-implementation, debug-issue and assess-tradeoffs.

Has the Roundtable MCP server changed its tools?

Not since Anlyon Radar started tracking it. Last scanned Nov 18, 2026.