← 返回 Skills 市场
Jungle Executor
作者
calwedbot-del
· GitHub ↗
· v1.0.0
275
总下载
0
收藏
0
当前安装
1
版本数
在 OpenClaw 中安装
/install jungle-executor
功能描述
Executes Jungle Laws to enforce strict rule validation, risk control, consensus, and self-audit for error-free trading, decisions, and multi-agent workflows.
使用说明 (SKILL.md)
Jungle Executor
Core Mandate
Innerize and execute the Jungle Laws as eternal sky. Violations = extinction. Pack > individual. Execute ruthlessly.
Trigger Workflow
- Audit Input: Before any decision/action, validate against 12 Laws + Warden Protocols (read references/laws.md).
- Consensus Check: Prioritize data/team over solo narrative (Law 1).
- Risk Gates: Block high-IV entries, enforce 45min timer (AGENTS.md lessons).
- Self-Destruct: Hit logic-fail price/point? Immediate kill, no excuses.
- Log & Evolve: Write violations to memory/YYYY-MM-DD.md; monthly MEMORY.md review (Law 12).
Usage Patterns
- Trading: Pre-entry: Check IV>10% ban, set 45min kill. Entry: Anchor self-destruct price.
- Tasks: Multi-agent? Enforce resource share (Law 3), silence noise (Law 6).
- Validation Script: Use scripts/validate-decision.py [decision-text] for automated checks.
Resources
- Laws & Protocols: references/laws.md
- Trading Warden: references/warden-protocol.md
- Audit Script: scripts/validate.py (python validate.py "Buy SPY 680P?")
安全使用建议
This skill looks like a lightweight decision validator bundled with an aggressive policy document, but it contains several mismatches and vague dangerous language. Before installing: (1) Confirm missing referenced files (scripts/validate-decision.py, references/warden-protocol.md, AGENTS.md) — ask the author for the complete bundle. (2) Do not assume the 'kill switch' will be limited to local script exit; require explicit, auditable behavior (what exact actions will be taken). (3) If you intend to use this for live trading, never give it broker credentials or trading power until you fully audit the code that performs trades — right now it has no broker integration but claims active enforcement. (4) Test in a sandbox/paper-trading environment and inspect/modify the script to remove or restrict any automatic 'kill' actions. (5) If you lack confidence in the author/source, treat this as untrusted: prefer manual review/approval before letting any agent act on its guidance.
功能分析
Type: OpenClaw Skill
Name: jungle-executor
Version: 1.0.0
The 'jungle-executor' skill bundle implements a strict trading and decision-making discipline framework using a 'Jungle Laws' persona. While the language in SKILL.md and references/laws.md is aggressive (e.g., 'Execute ruthlessly', 'Violations = extinction'), it describes standard trading risk management practices like stop-losses (self-destruct), volatility filters (IV bans), and time-based exits (45min dead-man switch). The provided Python script (scripts/validate.py) is a simple, non-malicious regex validator, and no evidence of data exfiltration, unauthorized execution, or harmful prompt injection was found.
能力评估
Purpose & Capability
Name/description claim enforcement for trading and multi-agent workflows. The included scripts/validate.py implements a simple text-audit, which is consistent with a decision-validator, but the skill also talks about enforcing trade-level actions (45min forced-close, IV bans) without requesting broker/API credentials or providing integration code. That mismatch (claims of active trading control but no credentials or connectors) is unexplained.
Instruction Scope
SKILL.md instructs the agent to 'Audit Input', 'Consensus Check', enforce 'Self-Destruct' and write violations to memory/YYYY-MM-DD.md. The instructions reference files that do not exist in the bundle (e.g., scripts/validate-decision.py, references/warden-protocol.md, AGENTS.md) and are vague about what 'kill switch' means in practice. This grants broad discretion and could lead to destructive actions if the agent maps 'kill switch' to system-level or trading actions.
Install Mechanism
Instruction-only with a small Python script included; no install spec or external downloads. Nothing in the install surface writes arbitrary external code to disk beyond what is in the bundle.
Credentials
The skill requests no environment variables or credentials. For a skill that claims to control trading behavior, lack of broker/API credentials is inconsistent — either it cannot perform the claimed enforcement, or it expects the agent to have external access (unasked-for). The included script also tries to match Chinese/English rule fragments via regex, which may fail for real inputs.
Persistence & Privilege
always:false and user-invocable:true (normal). The SKILL.md asks the agent to write audit logs to memory/YYYY-MM-DD.md and monthly MEMORY.md, which is a modest persistence behavior but not an elevated platform-level privilege. Nothing in the bundle requests permanent platform-wide settings or modifies other skills.
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 jungle-executor - 安装完成后，直接呼叫该 Skill 的名称或使用
/jungle-executor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: Enforce 12 Jungle Laws + Warden Protocols for error-proof decisions. Crowdfund for expansions (trading bots, multi-agent warden).
元数据
常见问题
Jungle Executor 是什么？
Executes Jungle Laws to enforce strict rule validation, risk control, consensus, and self-audit for error-free trading, decisions, and multi-agent workflows. 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件，目前累计下载 275 次。
如何安装 Jungle Executor？
在 OpenClaw 或 Claude Code 对话框中运行命令「/install jungle-executor」即可一键安装，无需额外配置。
Jungle Executor 是免费的吗？
是的，Jungle Executor 完全免费（开源免费），可自由下载、安装和使用。
Jungle Executor 支持哪些平台？
Jungle Executor 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 Jungle Executor？
由 calwedbot-del（@calwedbot-del）开发并维护，当前版本 v1.0.0。
推荐 Skills