📦 Chaos pivot — 混沌支点

v1.0.0

防止 LLM 陷入沉没成本谬误、强推已失效方案。当 agent 卡住、循环或连续失败时触发，强制进入 Popper 证伪时刻……

0· 631·0 当前·0 累计

by @manecharo (Manecharo)

AI模型访问自动化系统工具存储部署智能体

下载技能包

最后更新

2026/2/25

安全扫描

VirusTotal

无害

查看报告

OpenClaw

安全

high confidence

This is an instruction-only skill that coherently implements a structured “pivot” workflow for stuck agents; it requests no credentials, installs nothing, and stays within its stated purpose.

评估建议

This skill is internally coherent and low-risk: it only contains instructions for how an agent should recognize a dead end and try three very different alternatives. Before enabling it, confirm the agent's permitted actions when 'attempting' alternatives (e.g., whether it can call external APIs, run commands, or access files). Keep it user-invocable (do not set always:true), and consider adding operational guards: require explicit user approval before attempts that perform network calls or destr...

详细分析 ▾

✓ 用途与能力

Name, description, and runtime instructions all align: the skill teaches an agent to recognize dead-ends, produce three radically different alternatives, probe them, and pick or escalate. No extraneous permissions, binaries, or services are requested.

ℹ 指令范围

SKILL.md confines itself to internal reasoning steps (declare failure, produce 3 alternative lenses, probe each). It does give the agent broad discretion to 'attempt' alternatives — which could involve external actions depending on the agent's capabilities (API calls, CLI, or other tools the agent already has). The instructions do not explicitly tell the agent to read unrelated files, exfiltrate secrets, or contact unexpected endpoints, but the open-ended 'attempt' step means behavior depends on the agent's other tool permissions; consider restricting what the agent may do when probing alternatives.

✓ 安装机制

No install spec and no code files. Nothing is written to disk and there is no third-party code to fetch. Low installation risk.

✓ 凭证需求

The skill requires no environment variables, credentials, or config paths. There is no request for extra secrets or unrelated service tokens, which is proportional to the skill's described function.

✓ 持久化与权限

Flags: always:false and user-invocable:true (default). The skill does not request permanent inclusion or system-level configuration changes. It does not modify other skills or agent settings.

安全有层次，运行前请审查代码。

运行时依赖

无特殊依赖

版本

latestv1.0.02026/2/19

An anti-sunk-cost, chaos-injection mechanism into agentic reasoning. Falsification Moment (Phase 1) is the soul of the skill. Without naming why the approach failed and what assumption broke, any pivot is cosmetic. This is Popper applied directly — the failure must be honored as real information before being discarded. The 3 Lenses (Inversion, Analogy, Reduction) give the chaos structure. Fully random alternatives can waste cycles. Philosophically-guided alternatives — borrowed from how humans actually have breakthroughs — stay within the problem domain while being genuinely divergent. Inversion comes from TRIZ methodology, Analogy from how most scientific breakthroughs actually happen, and Reduction from the Unix philosophy and first-principles thinking. The Design Thinking loop gives it the iterative self-correcting structure you asked for. It doesn't just pivot once — it carries accumulated failure data forward so each loop is smarter than the last. The Escalation clause prevents infinite looping. After 3 full cycles, it surfaces the diagnosis to the user — which is itself a valuable output, not a failure.

● 无害

安装命令

点击复制

官方npx clawhub@latest install chaos-pivot

镜像加速npx clawhub@latest install chaos-pivot --registry https://cn.longxiaskill.com

数据来源：ClawHub ↗ · 中文优化：龙虾技能库