expert-council-router
Local, cost-aware multi-model expert orchestration for Pi and Codex MCP hosts: discovers callable Pi models, dynamically assembles semantic expert teams from billing policy and capability profiles, executes in isolated worktrees, and returns structured results to the Main Agent
Emerging. No concerning findings. Grades remain capped until the project builds reputation through adoption. Grades reflect the full trust pyramid: code, verification depth, and reputation. New projects cap at C until adoption is earned.
How we verified
Verified is a snapshot. Live keeps it current, and builds your track record.
⚡ Connect GitHub → continuous verification on every pushwhy connect →Who stands behind it
Labiey
Source: github_repo_search
Claim it to get a verified publisher badge, a free copy of our full audit findings, and direct contact for any high-priority issues we find. Or connect your repo for our deepest verification, Live Monitored: read-only, revoke anytime. What we access →
Install from
The grade above is for the source repository. Registries can serve a different version, so we mark the ones we were not able to read.
These names and descriptions are the publisher's own, read from the source code. We print them as written. Our assessment is the findings above, not this list.
expert_inspectInspect a compact summary of callable Pi resources. Request full detail only when exact model metadata is required.
expert_buildClassify a task and deterministically assemble a small semantic expert team. A current, complete, dated Main Agent model assessment is mandatory. Without composition or costPolicy the response lists up to 3 saved compositions plus an auto option; pass one back. Pass a saved composition name to restr…
expert_delegateStart one or up to eight bounded Pi expert assignments in the background and immediately return execution IDs. Prefer this over doing substantial multi-file investigation, implementation, review, or debugging inline whenever delegation saves Main Agent context or model quota; for a substantial task …
expert_waitAfter all other useful host work is finished, wait without polling for any or all background expert executions. Choose a finite timeoutMs from expected remaining task difficulty, then use expert_result for feedback.
expert_resultRetrieve completed expert feedback by execution ID, or report that the task is still running or unknown. Pass includeProgress while a task is still running to receive a bounded progress snapshot (last assistant output, elapsed time, files changed so far) for verification, handoff, or intervention de…
expert_abortDeliberately stop a running expert execution whose direction no longer matches expectations. The attempt is marked aborted and never retried or escalated, completed work such as a mutation worktree stays preserved until expert_cleanup, and the returned progress snapshot doubles as the handoff brief …
expert_feedbackRecord whether Main Agent verification accepted a completed expert result for local routing telemetry.
expert_cleanupRemove every retry/escalation worktree for an execution after its result has been integrated or rejected. Integrate changes from the result's filesChanged list (it already includes untracked new files) — never from `git diff HEAD` alone.
expert_escalateChoose a corrected retry, the next eligible model, or a bounded stop from structured failure evidence.
expert_statusReturn council plans, execution states, and local aggregate outcomes. view=summary (default) is a bounded running/recent/slots snapshot; view=full adds telemetry and the model assessment; view=running is only live executions.
expert_availability_resetClear runtime availability markers by scope: '*' (every model), a bare provider name (all its models), or an exact provider/id key. Use after a transient failure was misrecorded or a provider recovers before the marker TTL expires.
expert_verifyRun a bounded command on the plugin side inside a retained expert worktree (executionId) or a validated workspace, and return the real exit code and output tail. Turns 'the expert says it is green' into plugin-observed evidence.
expert_respondAnswer a running expert's pending interaction so its blocked turn continues in the same session. Headless hosts discover the open interaction by polling expert_status (view=running) or expert_result (includeProgress) and see a pendingInteraction. For a decision, set kind=decision and provide choice …
EXPERT_COUNCIL_MCP_TIMEOUT_MSbounded timeout for synchronous MCP operations, 30000 ms by default.EXPERT_COUNCIL_WORKSPACEdefault allowed workspace.All four hints declared on every tool
13/13 tools missing one or more hints — expert_inspect (missing: idempotentHint); expert_build (missing: idempotentHint); expert_delegate (missing: idempotentHint), +10 more. OpenAI's directory rejects tools where any of the four hints are missing or non-boolean.
For every tool, set all four hints (readOnlyHint, destructiveHint, idempotentHint, openWorldHint) to explicit true/false values that match the handler’s actual behaviour.
Tests exist
No test files found
Add tests that exercise each declared tool.
Domain consistency
npm scope @expert-council doesn't match GitHub owner labiey
Use the same org name across GitHub, npm, and your homepage so users can verify the publisher.
Claim the listing to review these findings one by one and send us a correction where you disagree, straight to the team. Claiming also means we tell you when the grade moves, and reach you first if we find anything urgent.
[](https://m8ven.ai/mcp/labiey/expert-council-router)?variant=verified to the badge URL.Vetting this one by hand? Tool Check is an MCP that scores other MCPs. Add it once and ask Claude, ChatGPT, or any MCP client to grade a server, surface CVEs, check the publisher, and suggest safer alternatives — before you install.
https://m8ven.ai/api/mcp/tool-check