Skip to content
verify mcp Beta VerifyMCP is currently in beta. If you notice any issues, get in touch and we’ll put it right.

io.github.omarkeshk/council-ai

REMOTE · MCP.COUNCIL-AI.APP · SCANNED SEP 20

Multi-LLM council: 25+ frontier models in parallel, consensus scoring, verdict-first code review.

0 this week 77 Trust /100

Recent critical change

Authorization (9 Aug 2026). See the changelog before you install this server.

Trust breakdown (7 categories)

How this component scores in each security and reliability category. Every signal is checked automatically against the live server, and we only credit what we can confirm. How we score → Why this is hard to score →

Endpoint Security57
Transport & Reachability100
Schema Quality & AI Usability71
  • AI-judged instruction clarity (excellent).Pass
  • Context-footprint check failed: tool/resource definitions use about 2128 tokens (~212/item across 10 items; 10 tools + 0 resources), over budget; trim descriptions and params. See how to fix → Fail
  • Usage-examples check failed: none of the tools include examples. See how to fix → Fail
Stability & Change Management100
  • No destabilizing schema changes in the last 30 days.Pass
Tool Coverage100
  • 100% of tools have a non-trivial description (not blank, and not just the tool's name).Pass
  • 100% of tool parameters carry a description.Pass
  • Structured output schemas are declared (100% of tools); any adoption earns full credit.Pass
Tool Safety88
  • No prompt-injection markers were found in the server instructions, tool names or descriptions we captured.Pass
  • 1 of 2 tool(s) whose name or description implies an irreversible operation declare an MCP destructiveHint annotation; "council_query" implies "send" and declares readOnlyHint instead, contradicting what its own name says it does. See how to fix → Partial
  • An AI judge read all 10 captured unit(s) of tool text and found none that tries to manipulate the model reading it.Pass
Capabilities100
  • Implements a supported MCP spec version (2025-11-25); the latest is 2026-07-28.Pass
Install

How do I install the io.github.omarkeshk/council-ai MCP server?

io.github.omarkeshk/council-ai is a hosted endpoint at https://mcp.council-ai.app/mcp, so there is nothing to install locally. Ready-made configuration for Claude, Cursor, VS Code, Codex and 5 more is on this page, copied from each client's own documentation.

remote · mcp.council-ai.app

# add to Claude Code
claude mcp add --transport http omarkeshk-council-ai 'https://mcp.council-ai.app/mcp'
// .cursor/mcp.json
{
  "mcpServers": {
    "omarkeshk-council-ai": {
      "url": "https://mcp.council-ai.app/mcp"
    }
  }
}
// .vscode/mcp.json
{
  "servers": {
    "omarkeshk-council-ai": {
      "type": "http",
      "url": "https://mcp.council-ai.app/mcp"
    }
  }
}
# ~/.codex/config.toml
[mcp_servers.omarkeshk-council-ai]
url = "https://mcp.council-ai.app/mcp"
// opencode.json
{
  "$schema": "https://opencode.ai/config.json",
  "mcp": {
    "omarkeshk-council-ai": {
      "type": "remote",
      "url": "https://mcp.council-ai.app/mcp",
      "enabled": true
    }
  }
}
# add to OpenClaw
openclaw mcp add omarkeshk-council-ai --url 'https://mcp.council-ai.app/mcp' --transport streamable-http
# ~/.hermes/config.yaml
mcp_servers:
  omarkeshk-council-ai:
    url: "https://mcp.council-ai.app/mcp"
// ~/.netclaw/config/netclaw.json
{
  "McpServers": {
    "omarkeshk-council-ai": {
      "Transport": "http",
      "Url": "https://mcp.council-ai.app/mcp"
    }
  }
}
# add to Vellum
assistant mcp add omarkeshk-council-ai -t streamable-http -u 'https://mcp.council-ai.app/mcp'
// mcp.json
{
  "mcpServers": {
    "omarkeshk-council-ai": {
      "type": "http",
      "url": "https://mcp.council-ai.app/mcp"
    }
  }
}

The mcpServers block is a cross-client convention. Remote transports vary, so check your client's docs.

Changelog

Every change we have recorded for this component, newest first. Security-relevant changes are always shown. ▲ marks a change for the better, ▼ a change for the worse; unmarked changes are neutral.

  • 20 Sept 26 0
    • This server's schema is too large to store in full, so we cannot compare its tools day to day functional
  • 19 Sept 26 0
    • This server's schema is too large to store in full, so we cannot compare its tools day to day functional
  • 18 Sept 26 0
    • This server's schema is too large to store in full, so we cannot compare its tools day to day functional
  • 17 Sept 26 0
    • This server's schema is too large to store in full, so we cannot compare its tools day to day functional
  • 16 Sept 26 0
    • This server's schema is too large to store in full, so we cannot compare its tools day to day functional
  • 15 Sept 26 0
    • This server's schema is too large to store in full, so we cannot compare its tools day to day functional
  • 14 Sept 26 0
    • This server's schema is too large to store in full, so we cannot compare its tools day to day functional
  • 13 Sept 26 0
    • This server's schema is too large to store in full, so we cannot compare its tools day to day functional
Diagnostics

Diagnostic detail from the automated scan of this channel: what the scanner observed at each step, so you can see exactly where a check passed or failed. It is informational only and never changes the trust score.

Captured 20 Sept 2026 · Probed https://mcp.council-ai.app/mcp

TLS valid

Negotiated TLS 1.3 with TLS_AES_128_GCM_SHA256 .

Subject Issuer Valid from Valid until Key Signature Serial
CN=mcp.council-ai.app CN=WE1,O=Google Trust Services,C=US 7 Sept 2026 6 Dec 2026 ECDSA 256 ECDSA-SHA256 3b2709b9897565d70ee6502e779acc05
SANs: mcp.council-ai.app
CN=WE1,O=Google Trust Services,C=US (CA) CN=GTS Root R4,O=Google Trust Services LLC,C=US 13 Dec 2023 20 Feb 2029 ECDSA 256 ECDSA-SHA384 7ff31977972c224a76155d13b6d685e3
CN=GTS Root R4,O=Google Trust Services LLC,C=US (CA) CN=GlobalSign Root CA,OU=Root CA,O=GlobalSign nv-sa,C=BE 15 Nov 2023 28 Jan 2028 ECDSA 384 SHA256-RSA 7fe530bf331343bedd821610493d8a1b

Background: What to check on a remote MCP endpoint →

DNSSEC insecure

Validation of mcp.council-ai.app. Not signed

Zone DS Keys Algorithms Outcome
. trust_anchor 20326, 38696 8, 8 Verified
app. present 23684 8 Verified
council-ai.app. absent Unsigned (proven) parent-signed NSEC/NSEC3 proves an unsigned delegation
Authentication No authorisation required

The endpoint answered without asking for a token. Anyone who knows the URL can reach it.

Result No authorisation required
HTTP status 200

Background: How OAuth 2.1 works in the 2026 MCP spec →

Transports 2 probes
Transport URL Outcome Status Location
streamable-http https://mcp.council-ai.app/mcp Verified 200
http (plaintext) http://mcp.council-ai.app/mcp HTTPS enforced 301 https://mcp.council-ai.app/mcp
MCP tools · 10 exposed · ~2,128 tokens

The tools this component advertises to a client, with an estimated token cost for each. Expand a tool to see its parameters and schema. The per-tool counts are indicative and are not scored directly; the schema's total context footprint is one signal in Schema Quality & AI Usability. A tool's description is untrusted text the model reads on every call, which is what makes this list a security surface and not just an inventory: how tool poisoning works →

Tool Tokens
council_models ~61

List the AI models available to the current user. Returns ID, provider, tier, context window, and capability flags (web search, vision, streaming). Use the IDs returned here as the `models` array argument to council_query / council_query_with_rag.

Input schema present but exposes no named parameters.

NameTypeReqDescription
modelsarrayyesModels available to this user, across all providers.
userPlanstringyesThe user's current subscription plan.

No examples provided.

council_query ~622

Send a prompt to a council of frontier AI models across 9 labs (Anthropic, OpenAI, Google, xAI, DeepSeek, Qwen, Mistral, Moonshot, z.ai). Returns each model's independent response plus a moderator-synthesized consensus answer with an agreement score and the key point of disagreement, when one exists. Use when a single-model answer might hallucinate or when verification across labs matters (research, contracts, architecture, legal, medical, code review). Adaptive-cost pattern: for a quick check, call with 2-3 models; if consensusScore comes back low (or keyDisagreement matters to the decision), escalate by re-running with more models — easy questions stay cheap, disputed ones get more compute. TIMEOUTS: a full council can run for minutes. If your MCP client has a fixed tool-call timeout, set async=true — you get a queryId back immediately and fetch the finished result with council_result. Call council_models first to pick specific model IDs, and council_usage to check remaining budget. Bills against the user's Council AI monthly budget.

NameTypeReqDescription
asyncbooleanWhen true, return {queryId, status:"running"} immediately instead of waiting for the council. Fetch the finished result with council_result. Use whenever your MCP client enforces a fixed tool-call ti…
modestringCross-model visibility. 'independent' (default): all models answer in parallel, blind to each other — best for unbiased comparison; the synthesis reconciles them. 'sequential': models answer one at a…
modelsarrayModel IDs to include (2-10). Defaults to a cross-lab flagship council. Use council_models to list available IDs. Example: ["claude-fable-5.1", "gpt-5.6-sol", "gemini-3.8-flash"]
promptstringyesThe question or task to send to the council. Include all needed context — every query is stateless. Example: "Is optimistic locking or a distributed lock the better fit for this checkout flow? [paste…
synthesisDirectivestringOptional instructions for the moderator that synthesizes the council's answers. The council models never see this — it only shapes the synthesis. Use for special modes, e.g. devil's advocate: "Argue…
synthesizebooleanWhen true (default), include the moderator synthesis + consensus score. Set false to get only the raw per-model responses (faster, cheaper).
NameTypeReqDescription
agreementTypeAgreement classification from the moderator's analysis. Null when synthesize=false.
budgetPercentUsednumber|nullThe user's monthly Council budget consumption after this query, as a percentage (0-100+). Use it to decide whether another query is affordable. Council never reports budget in dollar amounts — always…
consensusScorenumber|nullCross-model agreement score on the 0-100 scale. Null when synthesize=false; absent while an async query is still running.
consensusUnavailablebooleanTrue when the moderator's agreement verdict could not be parsed, so this run has no consensus score at all. Do not substitute a number — report agreement as unknown.
droppedModelIdsarrayRequested model IDs that were skipped (unknown or retired). Call council_models for current IDs.
durationMsnumberWall-clock duration of the council run (fan-out + synthesis), in milliseconds.
errorsarrayPer-model failures, if any. The synthesis covers the models that responded.
keyDisagreementstring|nullThe main point the models disagreed on, when one was identified. Often more decision-relevant than the consensus itself — inspect it before acting on a low consensusScore. Null on high agreement or w…
modelMappingobjectAnonymous reviewer label (M1..Mn) to model name mapping used during synthesis.
perModelResponsesobjectEach model's independent response, keyed by model name. Absent while an async query is still running.
queryIdstringServer-side ID of this council query. With async=true, pass it to council_result.
statusstringOnly present on async queries: "running" means poll council_result with the queryId.
synthesisstringModerator-synthesized consensus answer across all models. Empty when synthesize=false; absent while an async query is still running.
synthesisErrorstring|nullSet when moderator synthesis failed and the response degraded to raw per-model output.
timedOutModelIdsarrayCouncil members that failed by exceeding the per-model timeout. On a synchronous run (90s per model) this is recoverable: re-run the same query with async=true, which allows 300s. Slow reasoning mode…
totalTokensobjectAggregate token usage for this query across the whole council.
wasSyncRunbooleanTrue when this ran synchronously — the path with the tight per-model timeout.

No examples provided.

council_query_with_rag ~295

Like council_query, but first retrieves the most relevant passages from the user's personal Council RAG library (uploaded PDFs, Word docs, contracts, research papers, codebases) and injects them into every model's prompt. Use when the question is about content the user has uploaded — contract review, research synthesis across a paper library, code review against an architecture doc, etc. Ultra-tier only.

NameTypeReqDescription
asyncbooleanWhen true, retrieval runs immediately (sources are returned right away) but the council itself runs in the background — fetch the finished answer with council_result. Use when your MCP client enforce…
library_querystringOptional retrieval-specific query. If omitted, the main prompt is used as the retrieval query. Set this when the user's prompt has unrelated framing but the concrete documents to retrieve are describ…
modelsarrayModel IDs to include. Defaults to the user's preferred council.
project_idstringRestrict retrieval to a specific Council project workspace.
promptstringyesThe question or task to send to the council. Example: "Review this NDA draft against our standard playbook terms."
retrieval_limitnumberNumber of chunks to retrieve. Default: 8.
synthesizebooleanWhen true (default), include the moderator synthesis + consensus score.
NameTypeReqDescription
consensusScorenumber|nullCross-model agreement score on the 0-100 scale. Null when synthesize=false; absent while an async query is still running.
perModelResponsesobjectEach model's independent response, keyed by model name. Absent while an async query is still running.
queryIdstringPresent on async queries — pass it to council_result.
ragSkippedbooleanTrue when retrieval found nothing and the query fell through to a plain council_query.
retrievedChunksarrayyesLibrary chunks that were injected into every model's prompt. Empty when the library had no relevant content.
statusstringOnly present on async queries: "running" means poll council_result with the queryId.
synthesisstringModerator-synthesized consensus answer, grounded in the retrieved library chunks. Absent while an async query is still running.

No examples provided.

council_result ~97

Fetch the result of a council_query started with async=true. Returns status "running" while the council is still working (wait 30-60 seconds and call again), the full synthesis + per-model responses once complete, or status "failed" with the error. Polling is free — the council run itself was already billed by council_query.

NameTypeReqDescription
queryIdstringyesThe queryId returned by council_query when called with async=true.
NameTypeReqDescription
agreementTypeAgreement classification from the moderator's analysis.
budgetPercentUsednumber|nullMonthly Council budget consumption, as a percentage (0-100+). Never dollars.
consensusScorenumber|nullCross-model agreement score, 0-100. Null when the query ran with synthesize=false.
droppedModelIdsarrayRequested model IDs that were skipped (unknown or retired).
durationMsnumber|nullWall-clock duration of the council run, in milliseconds.
elapsedMsnumber|nullTime since the query started, while running.
errorstringFailure reason when status is "failed".
errorsarrayPer-model failures, if any. The synthesis covers the models that responded.
keyDisagreementstring|nullThe main point the models disagreed on, when one was identified.
modelMappingobjectAnonymous label (M1..Mn) to model ID mapping.
perModelResponsesobjectEach model's independent response, keyed by anonymous label (M1..Mn).
queryIdstringyesThe polled query ID.
statusstringyesrunning → poll again in 30-60s. complete → full result below. failed → see error.
synthesisstringModerator-synthesized consensus answer. Present when status is "complete".
synthesisErrorstring|nullSet when moderator synthesis failed and the response degraded to raw per-model output.
totalTokensobjectAggregate token usage across the whole council.

No examples provided.

council_review ~362

Multi-model code review. Sends a unified diff (or code snippet) to multiple frontier AI models from different labs in parallel, each acting as an independent reviewer with an explicit verdict + findings contract. Returns a verdict-first synthesis: overall SHIP/NO-SHIP, consensus score, confirmed findings (flagged by 2+ models), then dissents (single-model findings with reasoning), then each reviewer's verdict. Diffs are capped at 14,000 characters — split larger changes by file or hunk and call once per chunk. Bills against the user's Council AI monthly budget like any council query.

NameTypeReqDescription
asyncbooleanWhen true, return {queryId, status:"running"} immediately and fetch the finished review with council_result. Use when your MCP client enforces a fixed tool-call timeout, or with 4+ reviewers — async…
contextstringWhat the change is supposed to do: intent, constraints, related files or invariants reviewers cannot see from the diff alone. Improves signal, reduces false positives.
diffstringyesThe unified diff (git diff output) or code snippet to review. Max 14,000 characters — if your diff is larger, split it by file or hunk and review each part separately. Example: the output of `git dif…
focusstringReview focus: "bugs" (correctness), "security", "performance", "architecture", or "all" (default — full review, bugs prioritized).
modelsarrayModel IDs to include as reviewers (2-10). Defaults to the user's preferred council. Use council_models to list available IDs.
NameTypeReqDescription
consensusScorenumber|nullCross-reviewer agreement score on the 0-100 scale. Absent while an async review is still running.
errorsarrayPer-reviewer failures, if any.
focusstringyesThe review focus this synthesis was produced under.
modelMappingobjectAnonymous reviewer label (M1..Mn) to model name mapping — how dissents are attributed.
perModelResponsesobjectEach reviewer model's raw independent review, keyed by model name. Absent while an async review is still running.
queryIdstringServer-side ID of this review query. With async=true, pass it to council_result.
statusstringOnly present on async reviews: "running" means poll council_result with the queryId.
synthesisstringVerdict-first moderator synthesis: ## Council Verdict (SHIP/NO-SHIP), ## Confirmed Findings (2+ reviewers agree), ## Dissents, ## Reviewer Verdicts. Absent while an async review is still running.

No examples provided.

council_usage ~70

Return the user's current monthly cost-budget consumption (current spend, budget cap, percentage used, days until reset). Use to decide whether to warn the user before invoking another council_query, or to suggest using cheaper models. Per Council's rule: never show dollar amounts to the user in the response — use percentages.

Input schema present but exposes no named parameters.

NameTypeReqDescription
percentUsednumberPercentage of the monthly budget consumed, 0-100+.
periodEndsAtstringWhen the current monthly budget cycle resets, ISO 8601.
statusstringBudget status: ok, warning, or exceeded.

No examples provided.

library_delete ~122

Permanently delete a document from the user's Council RAG library — the record, every indexed chunk, AND the stored file are removed. This cannot be undone; re-adding the document requires uploading it again. Get document IDs from library_list. Use when the user asks to remove a document or when the 200-document library cap blocks an upload. Confirm with the user before deleting anything they did not explicitly name. No model call, no budget consumption.

NameTypeReqDescription
documentIdstringyesID of the document to delete, as returned by library_list or library_upload.
NameTypeReqDescription
deletedChunksnumberyesNumber of indexed chunks that were deleted along with the document.
successbooleanyesTrue when the document and all its data were removed.

No examples provided.

library_list ~47

List the documents in the user's Council RAG library. Returns id, filename, source type, ingestion status, chunk count, and upload date for each document. No retrieval, no budget consumption.

Input schema present but exposes no named parameters.

NameTypeReqDescription
documentsarrayyesEvery document in the user's library, newest first. Empty when the library has no documents.

No examples provided.

library_search ~155

Semantic search over the user's Council RAG library (uploaded PDFs, Word docs, contracts, research papers, codebases). Returns top-K chunks with source filename and page number. No model call, no budget consumption. Use to find direct quotes, check what the library contains, or scope a follow-up council_query_with_rag call.

NameTypeReqDescription
limitnumberNumber of chunks to return, 1-20. Default: 8.
project_idstringRestrict retrieval to a specific Council project workspace ID. Omit to search the whole library.
querystringyesSemantic retrieval query — describe the content you want, not keywords. Example: "termination and notice-period clauses in the vendor agreement"
NameTypeReqDescription
hitsarrayyesTop-K matching chunks, most relevant first. Empty when nothing matches.

No examples provided.

library_upload ~297

Upload a document into the user's Council RAG library so future council_query_with_rag and library_search calls can retrieve it. Accepts PDF, Word (docx), text, and markdown files as base64 — images are not supported. Max 10MB per file via MCP (the web library at https://council-ai.app/settings?tab=library takes up to 50MB); libraries hold up to 200 documents. Ingestion (chunking + embedding) runs in the background: the returned document starts in "pending" status — check library_list for it to reach "ready" before querying against it. No model call, no budget consumption.

NameTypeReqDescription
contentBase64stringyesThe file's raw bytes, base64-encoded (standard alphabet, no data: URI prefix). For a plain-text or markdown document, base64-encode the UTF-8 text.
filenamestringyesOriginal filename including extension — the extension drives type detection. Example: "vendor-contract-2026.pdf".
mimeTypestringMIME type, e.g. "application/pdf". Optional — the backend verifies the real type from the file's magic bytes regardless.
projectIdstringCouncil project workspace to file the document under. Omit for the general library.
tagsarrayFree-form tags to attach to the document, e.g. ["contracts", "2026"].
NameTypeReqDescription
documentobjectyesThe created document record, in "pending" status until background ingestion completes.

No examples provided.

Common questions

What is the io.github.omarkeshk/council-ai MCP server?

io.github.omarkeshk/council-ai is an MCP server listed in the public MCP registry as io.github.omarkeshk/council-ai. Multi-LLM council: 25+ frontier models in parallel, consensus scoring, verdict-first code review. This page covers its hosted endpoint (https://mcp.council-ai.app/mcp).

Is the io.github.omarkeshk/council-ai MCP server safe to use?

io.github.omarkeshk/council-ai scores 77 out of 100 on VerifyMCP. That is a record of what we were able to check automatically, not an endorsement. The category breakdown on this page shows every signal behind the number, including the ones we could not confirm.

What tools does the io.github.omarkeshk/council-ai MCP server expose?

io.github.omarkeshk/council-ai exposes 10 tools: council_query, council_result, council_query_with_rag, council_review, library_search, and 5 more. Their descriptions and schemas cost roughly 2,128 tokens of context every time the server is loaded.

Does the io.github.omarkeshk/council-ai MCP server require authentication?

No. We connected to io.github.omarkeshk/council-ai without credentials and it answered, so anything it exposes is reachable by anyone who knows the address.

Is the io.github.omarkeshk/council-ai MCP server still maintained?

io.github.omarkeshk/council-ai is still listed as active in the MCP registry. We last reached this channel on 20 September 2026. Those dates come from our own scans of the registry and the channel itself, not from anything the publisher announced.