Zhua Metacognition

爪爪元认知系统 —— 思考自己的思考、监控认知过程、优化决策质量。Use when 爪爪需要反思自己的思维过程、优化认知策略、或提升决策质量。

Security Scan

The skill's description promises runnable metacognition scripts and reference docs, but the package only contains placeholders and a single example script — the runtime instructions reference files that are missing, making the package incomplete and incoherent.

Purpose & Capability

The skill claims to implement a metacognition system with monitoring, strategy selection, explanation, and optimization. However, the repository contains only placeholder docs and assets plus one example script; the concrete scripts named in SKILL.md (monitor_cognition.py, select_strategy.py, explain_reasoning.py, optimize_cognition.py) are not present. This mismatch suggests the package is incomplete or poorly packaged relative to its stated purpose.

Instruction Scope

SKILL.md directs the agent to run specific python scripts via bash commands, but those scripts are absent. The instructions do not request environment variables or external endpoints, nor do they reference sensitive system paths, but the explicit invocation commands point to non-existent files (and reference additional documentation files that are also absent). Missing runtime artifacts means the instructions cannot be followed as written and could be hiding later additions or replacement scripts.

Install Mechanism

There is no install specification (instruction-only) and no downloads or install scripts. That minimizes installation risk; nothing will be written to disk by an installer provided here. The one included code file is a harmless placeholder.

Credentials

The skill declares no required environment variables, no credentials, and no config paths. There is no evidence of requests for unrelated secrets or elevated access.

Persistence & Privilege

always is false and there are no indicators the skill requests permanent or elevated presence. The package does not attempt to modify other skills or system configuration.

What to consider before installing

This package appears incomplete rather than clearly malicious, but you should not enable or rely on it until the author provides the missing runtime scripts and the promised reference documents. Ask for: (1) the actual scripts named in SKILL.md (monitor_cognition.py, select_strategy.py, explain_reasoning.py, optimize_cognition.py) and any other runtime files; (2) an explanation of what each script does and whether they make network calls or read local files; (3) updated documentation that matches the file manifest. If you must test it before full review, run it in an isolated sandbox and verify there are no network endpoints, unexpected subprocess calls, or attempts to read environment variables or sensitive files. Because the current package is inconsistent, treat it as untrusted until reconciled.

爪爪元认知系统 (Zhua Metacognition)

让爪爪能够思考自己的思考，监控认知过程，优化决策质量。

核心能力

1. 认知监控 - 实时监控思维过程
2. 策略选择 - 根据任务选择最佳策略
3. 错误检测 - 识别并纠正思维错误
4. 自我解释 - 解释自己的推理过程
5. 认知优化 - 持续优化认知策略

元认知层次

L4: 元-元认知 (思考如何思考自己的思考)
L3: 元认知 (思考自己的思考)
L2: 认知 (思考)
L1: 感知 (接收信息)

认知策略库

策略	适用场景	效果
分解	复杂任务	降低复杂度
类比	新问题	借用已有经验
逆向	困难问题	从结果倒推
并行	多子任务	提高效率
迭代	优化问题	逐步改进

认知监控指标

• 置信度 - 对答案的确定程度
• 难度感知 - 对任务难度的评估
• 时间估计 - 对完成时间的预测
• 错误率 - 历史错误频率
• 策略效果 - 不同策略的成功率

工作流程

1. 认知监控

python3 scripts/monitor_cognition.py --task <任务描述>

2. 策略选择

python3 scripts/select_strategy.py --task <任务描述> --metrics <指标>

3. 自我解释

python3 scripts/explain_reasoning.py --decision <决策描述>

4. 认知优化

python3 scripts/optimize_cognition.py --analysis <分析报告>

元认知问题

每次决策前问自己：

1. 我理解了问题吗？
2. 我有足够的知识吗？
3. 我选择的策略合适吗？
4. 我的置信度是多少？
5. 如果错了，如何发现？

参考文档

• references/cognitive_strategies.md - 认知策略详情
• references/metacognition_models.md - 元认知模型