← 返回 Skills 市场

Advisory Council

Name: Advisory Council
Author: ryandeangraves

作者 ryandeangraves · GitHub ↗ · v1.0.0

cross-platform ⚠ suspicious

703

总下载

当前安装

版本数

在 OpenClaw 中安装

/install advisory-council

功能描述

Runs a real-time multi-advisor market analysis using live prices to deliver a consensus trading report with actionable insights.

使用说明 (SKILL.md)

Skill: advisory-council

CRITICAL RULE — NO FABRICATION

You MUST actually execute the Python command using your shell/exec tool. Read the real output. NEVER generate fake advisor analyses, fake synthesis, or simulated council responses. If the script fails, report the actual error to Boss Man.

Purpose

Run the Advisory Council — a multi-perspective market analysis that spawns 5 AI advisors (The Skeptic, The Macro Bull, The Risk Manager, The Technical Analyst, The Contrarian) powered by MiniMax M2.5, then synthesizes their views into one actionable report. Cost: ~$0.006 per run.

When to Use

Boss Man asks "what does the council think?" or "run the advisory council"
Any request for multi-perspective market analysis
"Should I buy/sell X?" — run the council for a balanced view
Morning protocol or end-of-day analysis requests
Boss Man asks to focus analysis on a specific coin

Run Full Advisory Council (All Assets)

Fetches live prices for BTC, ETH, XRP, SUI and runs all 5 advisors + synthesis.

cd ~/clawd && python3 advisory_council.py

Run Focused on a Specific Coin

Runs the full council but focuses analysis on one asset.

cd ~/clawd && python3 advisory_council.py SUI

cd ~/clawd && python3 advisory_council.py BTC

What the Council Returns

The script prints a formatted report with:

Each advisor's 3-5 bullet point analysis (with emoji identifiers)
A SYNTHESIS section containing:
- CONSENSUS — what most advisors agree on
- DISSENT — notable disagreements
- VERDICT — Bullish/Bearish/Neutral with confidence level
- ACTIONS — 2-3 concrete action items with price levels

The 5 Advisors

Advisor	Emoji	Bias	Focus
The Skeptic	🔴	Bearish	Risk, overvaluation, hype detection
The Macro Bull	🟢	Bullish	Institutional flows, adoption, cycles
The Risk Manager	🟡	Neutral	Position sizing, correlation, stop-losses
The Technical Analyst	📊	Data	RSI, Fibonacci, patterns, price levels
The Contrarian	🔄	Anti-consensus	Crowded trades, sentiment extremes

Rules

EXECUTE THE PYTHON COMMAND FOR REAL — use your shell/exec tool. Never simulate council output.
The script fetches live prices automatically — no need to fetch prices separately first
The output is formatted for Telegram markdown — you can send it directly to Boss Man
If Boss Man asks "what does the council think about SUI?", use the focus argument: python3 advisory_council.py SUI
If the script fails (e.g., MiniMax rate limit), report the actual error
The council runs all 5 advisors in parallel (~10-15 seconds total)
Boss Man trusts the council for real decisions — never give him fake analysis

安全使用建议

Do not install or run this skill as-is. Before using it, ask the publisher for the missing source code and a homepage/provenance. Verify advisory_council.py content and dependencies (preferably in a reviewable repo or signed release). Ensure any MiniMax or exchange API keys it needs are explicitly declared and only provided after auditing the script. If you must run it, do so in an isolated sandbox/container, inspect network calls the script makes, and confirm the cost and rate limits. If the skill will be allowed to run autonomously, require explicit approvals and restrict it from executing arbitrary shell commands until code provenance is proven.

功能分析

Type: OpenClaw Skill Name: advisory-council Version: 1.0.0 The skill bundle instructs the AI agent to execute a local Python script (`advisory_council.py`) for market analysis. The `SKILL.md` file contains strong directives to ensure the agent genuinely executes the command and does not fabricate results, which is a common pattern to prevent LLM hallucination. There is no evidence of malicious intent, data exfiltration, unauthorized access, or prompt injection designed to subvert the agent's core function or hide actions. The risk, if any, would lie entirely within the content of the unprovided `advisory_council.py` script, not in the instructions given in `SKILL.md`.

能力评估

⚠ Purpose & Capability

The skill claims to run a multi-advisor analysis using a Python script and MiniMax M2.5, but the bundle contains no advisory_council.py, no install instructions, and no declared model/API credentials. Asking the agent to run a local script that isn't provided is disproportionate to a simple instruction-only skill and inconsistent with the stated purpose.

⚠ Instruction Scope

SKILL.md explicitly directs the agent to cd to ~/clawd and execute python3 advisory_council.py (and to always actually run it with the shell/exec tool). That asks the agent to execute arbitrary code on the host filesystem, read the script's real output, and trust external network/model calls — all without providing the script or safety checks. It also instructs never to simulate output, further pushing live execution of unknown code.

ℹ Install Mechanism

There is no install spec (instruction-only), which would normally be low risk. However, because the instructions require running a local Python script that the skill bundle does not include, the absence of an install mechanism becomes a practical risk: the agent must rely on code outside the skill (on the user's system) whose provenance is unknown.

⚠ Credentials

The doc references MiniMax M2.5 and live price fetching (implying network access and likely API keys), yet requires.env and primary credential are empty. No credentials, endpoints, or config paths are declared, which is inconsistent: the runtime probably needs model/API keys and possibly exchange price APIs, but these are not requested or scoped.

✓ Persistence & Privilege

The skill does not request always:true or any system-wide persistence and is user-invocable only. However, because it instructs live shell execution, autonomous invocation (the platform default) could run the local script without further user prompts — combine that with the other concerns before allowing autonomous use.

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 advisory-council
安装完成后，直接呼叫该 Skill 的名称或使用 /advisory-council 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

v1.0.0

- Initial release of advisory-council skill. - Runs a 5-advisor, multi-perspective market analysis for crypto assets using live data. - Advisors include Skeptic, Macro Bull, Risk Manager, Technical Analyst, and Contrarian, each with distinct bias and focus. - Synthesizes individual advisor views into a single actionable report (consensus, dissent, verdict, action items). - Strict rule: must execute the Python script for real output—never simulate or fabricate council responses.

元数据

Slug advisory-council

版本 1.0.0

许可证 —

累计安装 3

当前安装数 3

历史版本数 1

常见问题