← Back to Skills Marketplace
169
Downloads
0
Stars
0
Active Installs
7
Versions
Install in OpenClaw
/install zhuge-skill
Description
利用64卦推演和多数据源,提供足球比赛预测、经验自学习和多场决策支持的智能辅助工具。
Usage Guidance
This package appears internally consistent with its stated purpose. Before installing, note: (1) LLM relay/proxy mode (the 'relay' provider) will route prompts and keys through a third‑party server — avoid relay unless you trust that operator. (2) API_FOOTBALL_KEY and THE_ODDS_API_KEY (and any LLM keys you set) are used to call external services — only provide keys you control. (3) The repo writes local state (data/crystals_local.jsonl, experience.jsonl) and can optionally write a .env file but only after interactive confirmation. (4) If you want to be absolutely sure, run the recommended quick grep checks in SKILL.md (inspect scripts/sync.py to confirm it is pull‑only and scripts/share.py for no network calls). If you plan to use an LLM, prefer direct official endpoints (OpenAI / Anthropic / provider) and keep relay disabled.
Capability Analysis
Type: OpenClaw Skill
Name: zhuge-skill
Version: 1.1.1
The zhuge-skill is a football prediction engine that uses a combination of structured data from sports APIs (API-Football, The Odds API, Understat) and 64-hexagram reasoning. It supports 12 LLM providers for generating natural language commentary. The code demonstrates a high level of transparency and security awareness, including a dedicated PRIVACY.md, input sanitization in core/llm.py to mitigate prompt injection, and a pull-only synchronization model in scripts/sync.py to prevent unauthorized data upload. While it includes a 'relay' mode for LLM calls via third-party proxies, this is explicitly documented with a security warning. No evidence of malicious intent, data exfiltration, or unauthorized execution was found.
Capability Tags
Capability Assessment
Purpose & Capability
Name/description (64‑hexagram football predictions + experience crystallization) line up with the shipped adapters (API‑Football, The Odds, Understat), core modules (hexagram, yao, crystallizer) and scripts (predict, backfill, crystallize). No unrelated services (e.g., cloud provider credentials) are required.
Instruction Scope
SKILL.md instructs the agent to run local python scripts (predict, batch, backfill, sync pull, etc.). It documents which network calls occur (data APIs and chosen LLM provider). The guidance to grep/audit a few files is transparent. The instructions do not direct reading or transmitting unrelated system secrets.
Install Mechanism
No install spec is provided (instruction/code bundle only), so nothing is downloaded or executed by an automated installer. The code uses standard Python runtime and requests; installation is manual (pip) and limited to repository files.
Credentials
Many optional environment variables (12 LLM providers + two data APIs) are declared but all are optional and directly justify their purpose (data or LLM access). The skill does not require unrelated secrets. Be aware that providing an LLM key grants the skill the ability to send prompts to that provider.
Persistence & Privilege
always is false. The skill writes only to files within its own directory (data/*.jsonl, exports/, optional .env written only after explicit wizard confirmation). It does not alter system PATH or other skills' configs.
How to Use
- Make sure OpenClaw is installed (local or Docker)
- Run the install command in chat:
/install zhuge-skill - After installation, invoke the skill by name or use
/zhuge-skill - Provide required inputs per the skill's parameter spec and get structured output
Version History
v1.1.1
v1.1.1: 新手坑修复 - LLM_PROVIDER 必写, .env.example 补完整, sync URL 修正
v1.1.0
v1.1.0: 品牌视觉定稿 · brand.png v15 · 深海军蓝孔明×进化大脑
v1.0.4
UX + tooling batch: error_log, crystal v2 (first_seen/last_seen/recurrence_count), predict shows recurrence, stats --crystals, English README, cross-agent snippets, CHANGELOG.md
v1.0.3
Visual demo: embed 推演实录 screenshot in README (no code changes)
v1.0.2
Doc consistency fix: align SKILL.md with PRIVACY.md and actual code (no code changes)
v1.0.1
v1.0.1 透明度升级: (1) 新增省token核心卖点说明 (2) PRIVACY.md隐私合约 (3) sync.py重构为单向消费架构(仅pull) (4) llm.py relay模式运行时警告 (5) 完整环境变量清单
v1.0.0
Initial release: 5-source data fusion + 64-hexagram reasoning + experience crystallization. Part of the TaijiOS ecosystem.
Metadata
Frequently Asked Questions
What is Zhuge Skill?
利用64卦推演和多数据源,提供足球比赛预测、经验自学习和多场决策支持的智能辅助工具。 It is an AI Agent Skill for Claude Code / OpenClaw, with 169 downloads so far.
How do I install Zhuge Skill?
Run "/install zhuge-skill" in the OpenClaw or Claude Code chat to install it in one step — no extra setup required.
Is Zhuge Skill free?
Yes, Zhuge Skill is completely free, licensed under MIT-0. You can download, install and use it at no cost.
Which platforms does Zhuge Skill support?
Zhuge Skill is cross-platform and runs anywhere OpenClaw / Claude Code is available (cross-platform).
Who created Zhuge Skill?
It is built and maintained by 小九 (@yangfei222666-9); the current version is v1.1.1.
More Skills