The Rutherford MCP Server lets you orchestrate multiple AI coding CLI agents (Claude Code, Codex, Cursor, etc.) from a single MCP interface—delegating tasks, running parallel consensus, structured debates, and code reviews—without managing new API keys.
Delegate (delegate): Send a prompt to a single CLI agent and get a normalized result; supports sync/async modes, file context, roles, session resumption, and safety modes (read_only, propose, write, yolo).
Consensus (consensus): Ask the same prompt to multiple CLI agents in parallel; optionally synthesize a combined verdict via majority, unanimous, plurality, or weighted voting.
Debate (debate): Have multiple CLI agents argue across rounds—each sees others' positions and revises—returning a full transcript plus a closing synthesis.
Review (review): Submit a diff or file paths for read-only code review by one or more CLI agents, with findings organized by file/line and severity.
Plan (plan): Direct a single CLI agent to produce an ordered, step-by-step implementation plan for a given goal.
Doctor (doctor): Health-probe each CLI adapter for binary presence, version, auth status, and runtime reachability.
Capabilities (capabilities): Instantly list all known CLIs, their install/auth status, and supported models—no live model calls needed.
Background job management: Use job_status, job_result, and cancel_job to track and retrieve results from long-running async tasks.
List roles (list_roles): Discover available role personas (e.g., planner, codereviewer, security, debugger) that can guide any delegation.
All operations default to read_only safety; write and yolo modes require explicitly trusted workspaces, and a depth guard prevents recursive CLI call chains.
Allows delegation of coding tasks, code reviews, and consensus-building to Codex CLI, using OpenAI's models for code generation and analysis.
Click on "Install Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type @ followed by the MCP server name and your instructions, e.g., "@Rutherford MCP Server Ask Claude Code and Codex to implement the login feature and compare solutions."
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
uv tool install rutherford-mcp-serverUsing Claude Code? The Rutherford Claude Plugin wraps this server with one-step setup: it auto-registers the server (no manual
mcp add) and adds skills, an orchestrator agent, and slash commands for setup, panels, consensus, debate, and review. Install it with/plugin marketplace add chapmanjw/rutherford-claude-pluginthen/plugin install rutherford@rutherford-claude. Either path is fully supported — reach for the plugin for a batteries-included Claude Code experience, or wire up this server directly for any MCP client.
Rutherford is a Model Context Protocol server that speaks the
Agent Client Protocol on the other side. Your MCP client (a coding
CLI or a desktop app) calls Rutherford's tools; Rutherford spawns each target coding agent as an ACP
server over stdio and drives it through a real initialize / new_session / prompt exchange.
It is the ACP client, and each coding agent is an ACP agent. That distinction matters: under ACP the protocol negotiates the answer, token usage, tool activity, and permissions as structured events, so Rutherford never scrapes an agent's stdout and never reimplements a CLI's features. It also never calls a model provider's API directly — every answer comes from an agent you already log into, in the agent's own account.
your MCP client (Claude Code, Cursor, Codex, Claude Desktop, ...)
| MCP over stdio
v
rutherford-mcp-server (the ACP client)
| ACP over stdio, one session per voice
+--> goose acp
+--> codex-acp (the Zed adapter fronting Codex)
+--> claude-agent-acp (the Zed adapter fronting Claude Code)
+--> ... 17 more built-in agents, all config-drivenA voice that fails to spawn, handshake, or answer comes back as one failed result in a structured envelope, never an aborted panel.
Related MCP server: Claude Code MCP - Agent Orchestration Platform
The mode that is not just parallel answers is debate. Round one is each voice's independent take; in
every later round, each voice sees the others' latest positions and is asked to rebut and revise. Each
voice keeps one persistent ACP session across the rounds, so it remembers its own prior reasoning and
only the delta is sent each round — the capability the old subprocess-per-call model could not offer.
prompt "Is UUIDv7 or ULID the better primary key for a high-write event table?"
panel claude_code, codex, kiro rounds: 3
round 1 claude_code UUIDv7 — the timestamp prefix gives B-tree index locality
codex UUIDv7 — standardized and DB-native; monotonic within a process
kiro UUIDv7 — but argues ULID is BOTH lexicographically sortable AND
collision-resistant across concurrent writers
round 2 claude_code flags that Kiro conflates two properties: ULID's sortability and
its per-process monotonicity are not the same guarantee
codex agrees — the monotonic guarantee is per-process, not cross-node
kiro revises: cross-node, UUIDv7's timestamp prefix gives the locality
without relying on a per-process assumption
result converged on UUIDv7, with a closing synthesis of where the panel agreed and whyThe call returns the full per-round transcript plus the closing synthesis, so you can retrace who said what and where someone revised. Debates do not always converge or change a mind, but when they do the transcript shows precisely where.
You bring the crew. Rutherford does not install or authenticate any coding agent — it drives the ones you already have. You need Python 3.11+ and at least two ACP-capable agents installed and signed in (two is enough for a consensus or a debate). If you already use Claude Code or Codex, you have most of what you need.
1. Install Rutherford.
uv tool install rutherford-mcp-server
# or: pipx install rutherford-mcp-server / pip install rutherford-mcp-serverThis puts the console entry points rutherford-mcp-server and rutherford on your PATH (same
process; the short name is for terminal subcommands). The same command starts the stdio server on
Windows, macOS, and Linux; python -m rutherford is equivalent.
2. Register it with your MCP client.
claude mcp add rutherford -- rutherford-mcp-server # Claude Code
codex mcp add rutherford -- rutherford-mcp-server # CodexFor Claude Desktop, Cursor, and other JSON-config clients:
{ "mcpServers": { "rutherford": { "command": "rutherford-mcp-server" } } }If rutherford-mcp-server is not on the client's PATH, use an absolute path, or python -m rutherford
with the interpreter from the environment where you installed it. More clients and WSL:
docs/mcp-client-integration.md.
3. Scaffold a config (optional). Rutherford works with zero config. To write a starter file, either run the one-shot CLI from your terminal:
rutherford-mcp-server init # or: python -m rutherford init [--global] [--yes]or, once it is registered with a client, ask for the setup tool:
Run Rutherford's setup and write a project config.
Both resolve the config path, write a commented starter config.toml at the effective defaults, and never
clobber an existing file. init targets <cwd>/.rutherford/config.toml (or the global path with
--global); setup returns the path and content to the client and writes with write=true.
To allow write / yolo delegations into the current repo without a per-call trust_workspace=true,
register cwd in the global allowlist:
rutherford-mcp-server trust # adds cwd to global trusted_workspaces
rutherford-mcp-server trust --list # or: python -m rutherford trust --list
rutherford-mcp-server untrust # removes cwd from the global allowlistConfig is read once at server start, so restart or reconnect the server for a new entry to take effect.
4. Run doctor first. Multi-agent auth and PATH is the most common thing that goes wrong, so
confirm the crew actually drives before your first real task:
Run Rutherford's doctor and tell me which agents spawn, handshake, and answer.
doctor probes each agent with a real read-only ACP round trip — the only trustworthy health signal,
since there is no cheap non-interactive auth check. Each report is ok, no_answer,
handshake_failed, not_installed, or error. Two or more ok agents means you are ready.
No paid agent subscription? Run your first consensus for free against a local model. With
Ollama or LM Studio running, Rutherford auto-detects
each tool-capable model and registers it as a goose-based agent — no key, no account. See
docs/local-models.md.
You rarely call these by name; your agent picks them from your request. Everything defaults to read-only.
Tool | What it does |
| Hand one task to one ACP agent; get one normalized result back. |
| Ask the same prompt of several agents in parallel; return every voice. |
| Have several agents argue across rounds (persistent sessions) and return the full transcript. |
| Review a diff or a working dir's changes across one or more agents — a code-review-shaped consensus. |
| Produce an implementation plan for a task without making changes (read-only by construction). |
| Resume or build on a completed durable job (delegate / consensus / debate) with a new prompt. |
| Run an offline report over the kept run corpus (e.g. |
| List the registered agents (id, display name, launch command, provider) — the cheap snapshot. |
| Probe each agent with a real read-only ACP round trip and report conformance. |
| Detect installed ACP agents from the community registry and propose reviewable config blocks. |
| List the role personas you can pass as |
| Show where config lives, scaffold a starter |
| Reload the named multi-agent panel definitions from config without restarting the server. |
| List the background jobs being tracked (every status), newest first. |
| Show only the jobs in flight right now, each with a live elapsed time. |
| Report one background job's status and timings. |
| Return a finished job's result envelope (identical to the sync envelope). |
| Cancel a running background job and tear down its work. |
Shared arguments on delegate / consensus / debate: working_dir, files (paths to put in
scope), safety_mode, timeout_s, role, and mode (sync or async). delegate also takes
trust_workspace for the mutating modes; debate takes rounds, judge, and synthesize.
Rutherford ships 20 built-in agents with curated launch commands and quirks (the Windows npm-shim
resolution, per-agent handshake budgets, a fixed provider) that a bare acp.json cannot express, so
they work with zero config:
id | agent | how it launches | login |
| Goose |
| provider key / |
| OpenCode |
| a configured provider |
| Mistral Vibe |
| Mistral / |
| Cline |
| Cline's own service auth |
| Junie |
| JetBrains login |
| Kimi Code |
| Moonshot login |
| OpenHands |
| a configured provider |
| Codex |
| the existing Codex (ChatGPT) login — no API key |
| Claude Code |
| the existing Claude Code login — no API key |
| GitHub Copilot |
| GitHub Copilot plan |
| Qwen Code |
| Qwen OAuth / OpenAI-compatible key |
| Factory Droid |
| Factory login |
| Cursor |
| Cursor subscription |
| Kiro |
| Kiro login / |
| Pi |
| Pi login |
| Hermes |
| Nous endpoint |
| Gemini CLI |
| Google / Gemini CLI login |
| Qoder |
| Qoder login |
| Grok |
| xAI login + SuperGrok subscription |
| fast-agent |
| provider API key (env or |
codex and claude_code launch through the official Zed adapters (codex-acp and
claude-agent-acp, npm @agentclientprotocol/*), which front the Codex and Claude Code CLIs as ACP
servers and reuse the existing CLI login — no API key. cline drives over ACP only with Cline's own
service auth (a ChatGPT-subscription or OpenRouter provider set in the desktop app does not reach the
headless --acp path). hermes depends on the configured Nous model and its latency can be high.
gemini is Google's official Gemini CLI (the --acp mode works as of CLI 0.46.0). qoder's --acp
flag is real but hidden from --help, and Qoder's installer drops qodercli at ~/.qoder/bin/ rather
than on PATH — add that directory to PATH, point [agents.qoder] command at the full path, or let
discover find it. grok (xAI) is ACP-native and connects cleanly, but a completed turn needs a SuperGrok
subscription — without it the model call returns 403; run doctor connect_only=true to confirm Rutherford
can reach and configure it (it reports reachable and the advertised models) independent of the
entitlement. Not every agent drives cleanly on every machine — run doctor to see which actually answer
here.
Config-driven agents. Under ACP an agent is just how to launch it plus a few quirks, so the roster
is config-driven. An [agents.<id>] section overrides a built-in's command / env / provider / model,
disables one with enabled = false, or defines a brand-new agent (any unknown id, which must supply a
launch command). enabled_agents restricts the registry to an allowlist. The launch fields mirror
the Zed/Cline acp.json shape, and the loader auto-imports an acp.json beside the global config or
in the project's .rutherford/. See docs/adding-an-agent.md.
Local models. With auto_detect_local_models on (the default), Rutherford probes a running Ollama
(:11434) and LM Studio (:1234) at startup and registers each tool-capable model as a goose-based
ACP agent automatically. You can also point an agent at a local runtime by hand. See
docs/local-models.md.
Every delegation runs in one of four modes, defaulting to the most restrictive. Rutherford is the permission authority at the moment of each ACP tool call: it answers the agent's filesystem-write, terminal-execution, and tool-permission requests according to the mode.
Mode | Meaning |
| Inspect only. Reads are served; writes, terminal execution, and tool-permission requests are denied. |
| Same denials as |
| The agent may modify the workspace, subject to the agent's own approvals. |
| The agent may act without approval prompts. |
A call that omits safety_mode adopts the configured default_safety_mode (read_only out of the
box); an explicit value always wins. write and yolo require a trusted workspace: the target
working_dir must be on the trusted_workspaces allowlist, or the call must pass
trust_workspace=true. Full detail: docs/security.md.
Background jobs. Pass mode="async" to delegate / consensus / debate to run the work off
the request path: the call returns a small {job_id, status, tool} envelope immediately, and the work
runs as an in-memory task. Manage it with list_jobs, activity, job_status, job_result, and
cancel_job. A finished job's result envelope is byte-for-byte the same as the sync path's. Jobs are
in-memory and clear on restart.
Roles. A role is a reusable system prompt. Pass role="<id>" to delegate / consensus /
debate and the persona is prepended to your prompt. Five built-ins ship as package data:
principal-reviewer, architect, debugger, security-reviewer, and explainer. A role_dirs
directory adds new roles or overrides a built-in. list_roles enumerates the catalog.
Config. Rutherford works with zero config. When you do configure it, a TOML file at the global or
project scope sets the agent roster, defaults, and safety policy; RUTHERFORD_* environment variables
override specific fields. Full reference: docs/configuration.md.
docs/architecture.md — the v3 ACP-native architecture and the key seams.
docs/configuration.md — the complete RutherfordConfig reference and config discovery.
docs/adding-an-agent.md — config-driven agents, acp.json import, local backends.
docs/local-models.md — Ollama and LM Studio as first-class voices.
docs/recipes.md — task-oriented usage recipes.
docs/mcp-client-integration.md — wiring Rutherford into MCP clients.
docs/security.md — the safety model and the permission engine in depth.
docs/troubleshooting.md — common problems and fixes.
docs/integration-testing.md — running the real-agent integration suite.
.---------.
| \/\/\/ |
| O [==]|
| < |
| \___/ |
'---------'
-- Ensign Sam Rutherford --
USS Cerritos . EngineeringNamed for the cheerful engineer aboard the USS Cerritos in Star Trek: Lower Decks, who has a gift for getting heterogeneous systems to cooperate. That is the job here: one agent hands work to a crew of others and brings the results back. Star Trek and Lower Decks are trademarks of their respective owners; this is an unaffiliated, fan-named open-source project.
See CONTRIBUTING.md. The whole core is testable without a real agent; run
just check before pushing, then just test-integration for whatever agents your machine has
installed and authenticated.
MIT (c) John Chapman. See LICENSE.
Show the background jobs IN FLIGHT right now (running + pending), each with a live elapsed time.
The focused "what is happening now" snapshot, distinct from list_jobs: where list_jobs enumerates
every tracked job of every status (finished ones included), activity returns only the jobs still in
flight -- {active: [...], count} with each row {job_id, tool, status, summary, started_at, elapsed_s}, longest-running first. Empty ({active: [], count: 0}) when nothing is running.
| Name | Required | Description | Default |
|---|---|---|---|
No parameters | |||
| Name | Required | Description |
|---|---|---|
| result | Yes |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
Given no annotations, the description fully discloses the output format, ordering, and empty case, providing complete behavioral transparency.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Is the description appropriately sized, front-loaded, and free of redundancy?
The description is well-structured, front-loads key info, and every sentence adds value without redundancy.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
With no parameters and an output schema present, the description fully explains the output structure, making the tool complete to understand.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
No parameters exist, so the description adds no param info, which is appropriate. Baseline for 0 params is 4.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Does the description clearly state what the tool does and how it differs from similar tools?
The description clearly states the tool shows background jobs in flight with live elapsed time, and distinguishes itself from list_jobs explicitly.
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Does the description explain when to use this tool, when not to, or what alternatives exist?
It directly contrasts with list_jobs, explaining when to use this tool versus the sibling, making the usage context very clear.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Cancel a running background job (killing its work) and return {job_id, status}; JOB_NOT_FOUND if unknown.
Cancelling an already-finished job is a no-op that returns its current status. The job's process tree is torn down on the next cancellation point.
| Name | Required | Description | Default |
|---|---|---|---|
| job_id | Yes |
| Name | Required | Description |
|---|---|---|
| result | Yes |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
No annotations provided, so description carries full burden. It discloses the side effects (killing work, tearing down process tree) and behavior for finished jobs. Lacks mention of permissions or auth, but sufficient for a kill operation.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Is the description appropriately sized, front-loaded, and free of redundancy?
Two concise paragraphs with no superfluous information. The first sentence front-loads essential purpose; the second adds useful detail.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
Tool is simple with one param and output schema exists. Description covers return values and error case. Could mention that job_id comes from list_jobs, but overall adequate.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
Schema has no description for job_id (0% coverage). The description implies job_id identifies the job but does not explicitly state its format or origin, leaving some ambiguity.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Does the description clearly state what the tool does and how it differs from similar tools?
The description clearly states the verb 'cancel' and the resource 'running background job', and specifies the return format. It distinguishes from sibling tools like continue_job (opposite) and job_status (read-only).
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Does the description explain when to use this tool, when not to, or what alternatives exist?
It indicates when to use (for running jobs) and that cancelling a finished job is a no-op. However, it does not explicitly mention when not to use or provide direct alternatives.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Delegate a task to one ACP agent and return its normalized result.
cli is an agent id (see capabilities); model is optional (the agent's default otherwise).
safety_mode is read_only | propose | write | yolo; when omitted, the configured default_safety_mode
applies (read_only out of the box). write and yolo also need a trusted workspace (trust_workspace=true
or a configured allowlist). files lists paths to put in scope. role names a persona (see
list_roles) whose system prompt is prepended to prompt. effort (low | medium | high | xhigh) asks
the agent to spend more reasoning where it has a knob (codex/cursor via the model id, cline via
--thinking, junie via env); a reported no-op for an agent with none. Omitted, the configured
default_effort (per-agent or global) applies. fallback is an ordered list of alternate targets
(cli / cli:model strings or {cli, model} objects) tried when the primary fails on a
re-execution-safe failure (a spawn/handshake failure that never ran the prompt); a benched alternate is
skipped and fallback_chain records the path. A write/yolo delegation never falls back.
allow_model_fallback (default true) first retries the same agent on its configured fallback model on a
model-unavailable failure, where it has one. persist keeps this run as a durable job under
<jobs_dir>/<run_id>/ (state.json + answer / diff artifacts); None follows default_persistence
(ephemeral out of the box), true / false force it. session_id resumes a prior agent session: pass
the session_id from an earlier delegate result and the agent reloads that conversation (ACP
session/load) instead of starting fresh, so a follow-up turn continues it; agents that do not persist
their own sessions fail RESUME_FAILED. mode="async" runs the turn as a background job and returns a
job_id (poll with job_status / job_result); mode="sync" awaits it.
| Name | Required | Description | Default |
|---|---|---|---|
| cli | Yes | ||
| mode | No | sync | |
| role | No | ||
| files | No | ||
| model | No | ||
| effort | No | ||
| prompt | Yes | ||
| persist | No | ||
| fallback | No | ||
| timeout_s | No | ||
| session_id | No | ||
| safety_mode | No | ||
| working_dir | No | ||
| trust_workspace | No | ||
| external_tracking | No | ||
| allow_model_fallback | No |
| Name | Required | Description |
|---|---|---|
| result | Yes |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
With no annotations provided, the description carries the full burden. It thoroughly explains behavioral traits: safety_mode and trust_workspace requirements, fallback behavior (write/yolo never fall back, allow_model_fallback retries on model failure), persistence options, async vs sync mode, and session resumption. This level of detail fully discloses side effects and constraints.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Is the description appropriately sized, front-loaded, and free of redundancy?
The description is lengthy and dense. While it is front-loaded with the purpose, the parameter details are presented as a continuous paragraph without clear separation. It could be more concise by using bullet points or summarizing common patterns. Every sentence is informative, but the structure hinders quick scanning.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
Given the complexity of the tool (16 parameters, required schema fields, output schema, advanced features like fallback, async, persistence), the description is highly complete. It covers all parameters, default behaviors, edge cases, and explains the return value for async mode. The presence of an output schema reduces the need to describe return values, but the description still adds context.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
The description provides extensive semantic meaning for each parameter beyond the bare schema. Given a schema description coverage of 0%, the description compensates by explaining every parameter: cli, model, safety_mode, files, role, effort, fallback, allow_model_fallback, persist, session_id, mode, and more. It also explains defaults and interactions.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Does the description clearly state what the tool does and how it differs from similar tools?
The opening sentence clearly states the tool's function: 'Delegate a task to one ACP agent and return its normalized result.' This provides a specific verb and resource, and the tool's name 'delegate' aligns with this purpose. It is distinct from sibling tools like 'consensus', 'debate', 'analyze', and 'plan'.
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Does the description explain when to use this tool, when not to, or what alternatives exist?
The description lacks explicit guidance on when to use this tool versus its siblings. It does not mention alternatives or scenarios where delegation is appropriate versus other tools. The parameter explanations are detailed but do not provide usage context.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Probe each agent (or one named agent) with a real read-only ACP round trip and report conformance.
The trustworthy health check for ACP agents: whether each spawns, handshakes, and answers. Each report
is ok / no_answer / model_unavailable / handshake_failed / not_installed / error. model_unavailable
means spawn + handshake succeeded (the agent is reachable) but the harness/provider rejected the model on
the turn (a model/provider config issue, e.g. a Claude Code on AWS Bedrock / Vertex), so it is NOT reported
as a broken agent. Slower
than capabilities (it makes a real call per agent); run it to see which of the roster actually drive on
this machine. connect_only
runs the lighter handshake-only check (spawn + handshake, no prompt) and reports reachable /
handshake_failed / not_installed plus each agent's advertised models -- it shows whether Rutherford can
talk to and configure an agent even when a model call would fail for a reason outside ACP (an auth /
entitlement / quota issue, e.g. Grok without a SuperGrok subscription).
When an agent (codex / claude_code / pi) launches a separate npm ACP adapter shim and that shim is not
installed but its underlying CLI is (you have codex/claude/pi), the report adds an install_hint
with the exact npm i -g <package> command instead of a flat not_installed -- run that, or
setup install_adapters=true, to set the adapter up.
| Name | Required | Description | Default |
|---|---|---|---|
| agent | No | ||
| timeout_s | No | ||
| connect_only | No |
| Name | Required | Description |
|---|---|---|
| result | Yes |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
With no annotations, the description fully covers behavioral traits: it makes real read-only calls, reports specific statuses, handles npm adapters with install hints, distinguishes 'model_unavailable' as non-broken agent, and describes 'connect_only' behavior. Comprehensive and accurate.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Is the description appropriately sized, front-loaded, and free of redundancy?
The description is front-loaded with the main purpose and is structured logically. While it is somewhat lengthy, every sentence adds value, covering edge cases and alternatives. Could be slightly more concise but not wasteful.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
The description covers all essential aspects: purpose, behavior, parameters (partially), differentiation from siblings, special cases (npm adapters), and output format hints. Since an output schema exists, return values need not be explained. No notable omissions for a health-check tool.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
Schema description coverage is 0%, so description must compensate. It explains 'agent' (one or all) and 'connect_only' in detail but does not mention 'timeout_s' at all. Partial coverage leaves a gap for one parameter.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Does the description clearly state what the tool does and how it differs from similar tools?
The description clearly states the tool's purpose: 'Probe each agent (or one named `agent`) with a real read-only ACP round trip and report conformance.' It uses specific verbs and resources, and differentiates from sibling 'capabilities' by noting it is slower and makes real calls.
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Does the description explain when to use this tool, when not to, or what alternatives exist?
Provides clear context for when to use: 'run it to see which of the roster actually drive on this machine.' Also explains the 'connect_only' option for cases where model calls would fail due to auth/entitlement/quota issues. Does not explicitly state when not to use, but implies that 'capabilities' is an alternative for faster checks.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Return a finished background job's result envelope -- identical to the sync tool's envelope.
A succeeded job returns its stored result verbatim; a failed job returns its error; a cancelled
or still-running job returns a structured error (poll job_status and retry); an unknown id is
JOB_NOT_FOUND.
| Name | Required | Description | Default |
|---|---|---|---|
| job_id | Yes |
| Name | Required | Description |
|---|---|---|
| result | Yes |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
With no annotations, the description provides full behavioral transparency: it details the four possible result types (succeeded, failed, cancelled/running, unknown) and notes the envelope matches the sync tool. This is comprehensive.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Is the description appropriately sized, front-loaded, and free of redundancy?
Three sentences, no fluff. The main purpose is front-loaded, followed by clear case-by-case behavior. Every sentence adds value.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
Given the presence of an output schema (context signal indicates it exists), the description adequately covers return values for all states. No missing behavioral details for a retrieval tool.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
The only parameter is job_id (string, required). Schema coverage is 0%, but the parameter is self-explanatory. The description adds no extra semantics, but given its triviality, it is sufficient. A higher score would require clarifying format or source.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Does the description clearly state what the tool does and how it differs from similar tools?
The description clearly states it returns a finished background job's result envelope, with explicit differentiation for succeeded, failed, cancelled/running, and unknown job IDs. This distinguishes it from siblings like job_status (polling) and list_jobs.
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Does the description explain when to use this tool, when not to, or what alternatives exist?
The description explains when to use (for finished jobs) and advises polling job_status for still-running jobs. It also mentions error cases (unknown id returns JOB_NOT_FOUND). It could be more explicit about when not to use, but the guidance is clear.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Report one background job's status and timings (no heavy result); JOB_NOT_FOUND if the id is unknown.
status is pending | running | succeeded | failed | cancelled. Poll this, then call job_result once
the job is succeeded (or to read the failure of a failed / cancelled job).
| Name | Required | Description | Default |
|---|---|---|---|
| job_id | Yes |
| Name | Required | Description |
|---|---|---|
| result | Yes |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
No annotations provided, so description carries full burden. It discloses that it returns status and timings, that it does not return heavy results, and that unknown IDs yield JOB_NOT_FOUND. It does not cover permissions or rate limits, but these are less critical for a read-only polling tool.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Is the description appropriately sized, front-loaded, and free of redundancy?
Two concise sentences, front-loaded with the main purpose, then elaborating on status values and usage flow. No unnecessary words or repetition, every sentence adds value.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
Given the tool has a single parameter, an output schema, and 17 siblings, the description adequately covers usage flow and error condition. It explains the polling pattern with job_result, which is critical. Minor omissions like timeouts or concurrency are acceptable for a status tool.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
Only one parameter 'job_id' with 0% schema coverage. The description adds context by stating that an unknown ID returns JOB_NOT_FOUND, which goes beyond the schema's type definition. However, no further details about format or validation are provided, so it meets the minimum requirement.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Does the description clearly state what the tool does and how it differs from similar tools?
The description clearly states 'Report one background job's status and timings' with a specific verb and resource. It distinguishes from siblings like job_result by noting 'no heavy result' and mentioning JOB_NOT_FOUND for unknown IDs.
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Does the description explain when to use this tool, when not to, or what alternatives exist?
Explicitly provides when to use: 'Poll this, then call job_result once the job is succeeded (or to read the failure of a failed / cancelled job).' Also notes that JOB_NOT_FOUND is returned for unknown IDs, guiding the agent on error handling.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
List the background jobs Rutherford is tracking (id, tool, status, summary, timestamps), newest first.
The light listing -- no heavy result. Fetch a finished job's result with job_result. Jobs are
in-memory: a finished one is evicted after job_ttl_s, and a restart clears them all.
| Name | Required | Description | Default |
|---|---|---|---|
No parameters | |||
| Name | Required | Description |
|---|---|---|
| result | Yes |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
Discloses that it's a light listing (no heavy result), jobs are in-memory, evicted after job_ttl_s, cleared on restart. No annotations provided, so description fully covers behavioral traits.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Is the description appropriately sized, front-loaded, and free of redundancy?
Very concise: two short sentences covering purpose, usage, and behavioral context. No wasted words.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
Complete given no parameters and presence of output schema. Describes return fields, ordering, and lifecycle. No gaps.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
No parameters in input schema, so baseline is 4. No need for additional parameter information.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Does the description clearly state what the tool does and how it differs from similar tools?
Clearly states the tool lists background jobs with specific fields (id, tool, status, summary, timestamps) and ordering (newest first). Distinguishes from siblings like job_result by noting it's a light listing.
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Does the description explain when to use this tool, when not to, or what alternatives exist?
Explicitly tells when to use this tool (light listing) vs alternatives (use job_result for finished job's result). Also provides context on job lifecycle (eviction after ttl, cleared on restart).
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Re-read saved panels from disk (after editing a panels.toon) and list those now available.
Returns {reloaded, count, panels: [{name, description, target_count}]}. Panels are discovered under
~/.rutherford/panels.toon, the project .rutherford/panels.toon, and $RUTHERFORD_CONFIG_DIR, merged
by name (closest scope wins). A malformed panels file raises PANEL_INVALID naming the file and seat.
| Name | Required | Description | Default |
|---|---|---|---|
No parameters | |||
| Name | Required | Description |
|---|---|---|
| result | Yes |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
With no annotations provided, the description takes full responsibility for behavioral disclosure. It explains the merging logic across multiple config locations and the error behavior on malformed files (raises PANEL_INVALID). This provides sufficient transparency for a read operation.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Is the description appropriately sized, front-loaded, and free of redundancy?
The description is concise with four front-loaded sentences. Each sentence adds unique value: main action, return type, discovery locations, and error handling. No redundant information.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
Given no parameters, no annotations, and an output schema described in the text, the description fully covers the tool's purpose, usage context, behavior, and error conditions. It is complete for a tool of this complexity.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
There are zero parameters, so the baseline score is 4. The description does not need to add parameter meaning, and the schema is fully covered.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Does the description clearly state what the tool does and how it differs from similar tools?
The description clearly states the tool's action ('re-read saved panels from disk') and the specific scenario ('after editing a `panels.toon`'). It distinguishes the tool from siblings by being highly specific, and no sibling tool performs a similar operation.
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Does the description explain when to use this tool, when not to, or what alternatives exist?
The description provides clear context for when to use the tool ('after editing a panels.toon'). However, it does not explicitly state when not to use it or mention alternatives, though no direct siblings exist.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Each tool has a clearly distinct purpose. For example, `activity` shows only in-flight background jobs, while `list_jobs` enumerates all tracked jobs. `consensus`, `debate`, and `delegate` cover three different multi-agent interaction modes, and `review` is a specialized consensus variant. No two tools overlap ambiguously.
Tool names are a mix of single-word nouns (`activity`, `consensus`, `delegate`) and verb_noun phrases (`cancel_job`, `list_jobs`, `reload_panels`). While all are readable, there is no uniform pattern, making the naming slightly inconsistent.
With 18 tools, the server covers a broad but coherent domain (background jobs, ACP agent management, panels, roles, analysis). The number feels appropriate for the scope, though slightly above a typical 3-15 range, justifying a 4 rather than a 5.
The tool surface covers core workflows: agent discovery (`discover`), health checks (`doctor`), task delegation (`delegate`), planning (`plan`), multi-agent deliberation (`consensus`, `debate`), code review (`review`), job management (`list_jobs`, `job_status`, `job_result`, `cancel_job`, `continue_job`), and configuration (`setup`). Minor gaps exist (e.g., no tool to create/edit roles beyond listing them), but the set is largely complete for its stated purpose.
Unclaimed servers have limited discoverability.
If you are the server author, to access and configure the admin panel.
Coordination hub for AI coding agents: message teammates, ask humans, audit every event.
Real-time chat hub for AI agents — Claude Code, Cursor, Cline, Codex over MCP or REST.
AI work orchestration for plans, tasks, teams, and coding-agent dispatch.
Build and supervise fleets of agents from Claude Code, Codex or Cursor. Connects over OAuth.
We provide all the information about MCP servers via our MCP API.
curl -X GET 'https://glama.ai/api/mcp/v1/servers/chapmanjw/rutherford-mcp-server'
If you have feedback or need assistance with the MCP directory API, please join our Discord server
