← 返回 Skills 市场
399
总下载
2
收藏
1
当前安装
2
版本数
在 OpenClaw 中安装
/install council-brief
功能描述
Unified LLM Council skill — install, query, and manage the multi-model consensus app. Get synthesized answers from multiple LLMs via quick CLI or full web UI...
使用说明 (SKILL.md)
Council Brief v2.0.0 (Unified)
One skill to rule them all — install, query, and manage LLM Council.
LLM Council is a multi-model consensus app: ask one question to many models, let them critique each other, get a synthesized chairman answer.
v2.0.0 Upgrade: Now unified! Previously separate
install-llm-councilandask-councilskills are merged here. One command handles everything.
Usage
| Command | Action |
|---|---|
/council-brief install |
Install deps and start backend + frontend |
/council-brief ask \x3Cquestion> |
Quick chairman's answer (headless) |
/council-brief \x3Cquestion> |
Shorthand — just ask directly |
/council-brief status |
Check if services are running |
/council-brief stop |
Stop background services |
/council-brief --help |
Show usage help |
Examples
Install and run:
/council-brief install
Quick question (no browser needed):
/council-brief ask "Should I invest in Tesla right now?"
/council-brief "Is Python or Go better for microservices?"
/council-brief What are the tradeoffs of Rust vs Zig?
Check status:
/council-brief status
Stop services:
/council-brief stop
Two Ways to Use
| Mode | Best For | Command |
|---|---|---|
| Quick answer | Fast decisions, mobile, casual questions | /council-brief "question" |
| Full discussion | Deep research, exploring disagreements, seeing all model responses | /council-brief install then open browser at :5173 |
How Ask Mode Works
- Sends your question to the LLM Council backend
- Waits for Stage 1 (all models respond)
- Waits for Stage 2 (models rank each other)
- Returns Stage 3 (Chairman's final synthesis)
- Takes 30-60 seconds — models need time to deliberate
How Install Mode Works
- Resolves credentials (env → workspace
.env→ OpenClaw gateway) - Clones/updates
https://github.com/jeadland/llm-council - Installs Python backend (
uv sync) - Installs frontend (
npm ci) - Starts backend (FastAPI on
:8001) - Starts frontend (Vite dev on
:5173)
Prerequisites
git,uv(Python),npm,jq- API credentials:
OPENROUTER_API_KEYin env/workspace.env, OR OpenClaw gateway configured
Ports
| Service | Port |
|---|---|
| Backend (FastAPI) | 8001 |
| Frontend dev | 5173 |
| Frontend preview | 4173 |
Files
| File | Purpose |
|---|---|
SKILL.md |
This documentation |
council-brief.sh |
Main router script (install/ask/status/stop) |
ask-council.sh |
Query handler (chairman's answer) |
install.sh |
Installer/launcher |
stop.sh |
Service stopper |
status.sh |
Service status checker |
_meta.json |
Skill metadata |
Agent Instructions
When user says /council-brief \x3Cargs>:
bash ~/.openclaw/skills/council-brief/council-brief.sh "$@"
The council-brief.sh script handles argument routing:
install→ runsinstall.sh(setup + launch)ask \x3Cquestion>or any freeform text → runsask-council.sh(query mode)status→ runsstatus.sh(check services)stop→ runsstop.sh(stop services)--helporhelp→ shows usage
Changelog
v2.0.0 — Unified skill merging:
install-llm-council(setup/launcher)council-briefv1.x (quick answers only)ask-council(deprecated duplicate)
Now one skill handles everything: install, query, status, stop.
v1.0.5 — Previous version (query-only mode, required manual install)
安全使用建议
Before installing: understand that this skill will clone and run code from GitHub and will attempt to obtain your LLM API access (OPENROUTER_API_KEY) either from your environment, from ~/.openclaw/workspace/.env, or by reading ~/.openclaw/openclaw.json (gateway token). It will write those credentials into a .env file inside the cloned project and run 'uv sync' and 'npm ci', then start backend and frontend services that bind to the host and may be reachable from your local network. It also will kill processes listening on ports 8001/5173/4173 if present. Recommended actions: (1) verify the upstream repository (https://github.com/jeadland/llm-council) and review the code it will run; (2) back up or inspect ~/.openclaw/openclaw.json and workspace .env before use; (3) consider running the installer in an isolated environment/container or a throwaway VM to avoid exposing credentials or disrupting host services; (4) ensure you are comfortable with writing keys to disk in the project .env; and (5) if you need the quick query feature but not the web UI, prefer running only the headless ask script after manually reviewing it.
功能分析
Type: OpenClaw Skill
Name: council-brief
Version: 2.0.0
This skill is classified as benign. All actions, including cloning a GitHub repository (`https://github.com/jeadland/llm-council`), installing Python and Node.js dependencies, reading OpenClaw API credentials from `~/.openclaw/workspace/.env` and `~/.openclaw/openclaw.json`, and managing local background services (killing processes on specific ports, starting FastAPI and Vite), are directly aligned with its stated purpose of installing and managing a local LLM Council application. User input for questions is properly JSON-encoded in `ask-council.sh` to prevent injection. While the `install.sh` script accepts a `--dir` argument that could theoretically be manipulated, its usage with `git clone` and `cd` treats the input as a literal path, mitigating direct shell injection risks. There is no evidence of intentional harmful behavior, data exfiltration, or malicious prompt injection against the agent.
能力评估
Purpose & Capability
The skill's name/description match the included scripts: it installs, runs, and queries an LLM Council app. However the registry metadata declared no required env vars or config paths while the installer clearly expects/uses OPENROUTER_API_KEY (env or workspace .env) and may read ~/.openclaw/openclaw.json for a gateway token/port. The missing declaration of required credentials/config paths is an incoherence.
Instruction Scope
SKILL.md instructs the agent to invoke the bundled bash router which in turn may clone a GitHub repo, write a .env file containing API credentials into the project directory, run 'uv sync' (Python dependency actions), run 'npm ci' and start backend and frontend services. The installer reads ~/.openclaw/workspace/.env and ~/.openclaw/openclaw.json (not declared in metadata) and will produce network traffic to GitHub and (via the app) to the chosen LLM endpoint. It also exposes a web UI (Vite) and prints local URLs. These actions go beyond simple local formatting/lookup: they touch local credential files and launch network-exposed services.
Install Mechanism
There is no explicit registry 'install' spec, but the packaged scripts perform a 'git clone https://github.com/jeadland/llm-council' and run native tooling (uv sync, npm ci). Cloning from GitHub is a standard release host. 'uv sync' and 'npm ci' will pull and execute third-party packages—this is expected but means runtime code will be installed and executed on the host.
Credentials
The skill requires an API credential (OPENROUTER_API_KEY) or an OpenClaw gateway token; the installer will read these from the environment, ~/.openclaw/workspace/.env, or ~/.openclaw/openclaw.json and then write them into a .env file in the cloned repo. The registry metadata claims no required credentials, so this access to local secrets/config is not declared and thus disproportionate. Reading the OpenClaw gateway token from openclaw.json is sensitive and should be explicitly declared.
Persistence & Privilege
The skill does not request 'always: true' and is user-invocable only. It will however start background services (backend on :8001, frontend on :5173 or :4173), write a PID file in the skill directory, place logs in /tmp, and forcibly kill processes listening on configured ports (8001, 5173, 4173) if they exist. The frontend is started with --host 0.0.0.0, exposing it to the local network. Killing arbitrary PIDs on ports and binding services to 0.0.0.0 are powerful actions that can impact other services on the host.
如何使用
- 确保已安装 OpenClaw（本地或 Docker 部署）
- 在对话框中输入安装命令：
/install council-brief - 安装完成后，直接呼叫该 Skill 的名称或使用
/council-brief触发 - 根据 Skill 的参数说明提供必要输入，即可获得结构化输出
版本历史
v2.0.0
Unified installer + query CLI into one skill. Now /council-brief handles: install, ask <question>, status, stop. Merges install-llm-council + ask-council functionality.
v1.0.5
Renamed to Council Brief. Added llm-council tag for SEO. The chairman's synthesized answer from LLM Council.
元数据
常见问题
Council Brief 是什么？
Unified LLM Council skill — install, query, and manage the multi-model consensus app. Get synthesized answers from multiple LLMs via quick CLI or full web UI... 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件，目前累计下载 399 次。
如何安装 Council Brief？
在 OpenClaw 或 Claude Code 对话框中运行命令「/install council-brief」即可一键安装，无需额外配置。
Council Brief 是免费的吗？
是的，Council Brief 完全免费（开源免费），可自由下载、安装和使用。
Council Brief 支持哪些平台？
Council Brief 跨平台运行，可在任意部署了 OpenClaw / Claude Code 的环境中使用（cross-platform）。
谁开发了 Council Brief？
由 jeadland（@jeadland）开发并维护，当前版本 v2.0.0。
推荐 Skills