Thinking - 深度思考与规划
v1.1.0通过系统化元认知流程，深入解析任务、盘点技能、检索经验、规划路径与评估风险，提升复杂任务执行效果与效率。
⭐ 0· 94·0 current·0 all-time
MIT-0
Download zip
LicenseMIT-0 · Free to use, modify, and redistribute. No attribution required.
Security Scan
OpenClaw
Benign
high confidencePurpose & Capability
The name/description (meta-cognitive thinking and planning) align with the instructions: task analysis, skill inventory, memory retrieval, planning, and risk assessment. The skill references listing installed skills and searching agent memory/history — actions coherent with a planning/meta-skill.
Instruction Scope
SKILL.md instructs the agent to access agent-local resources (memory via memory_search, session history via sessions_history, and the skills list). Those are sensitive but appropriate for a thinking/meta-skill that must consult past context and available capabilities. The doc does not instruct reading unrelated system files or exfiltrating data externally. Examples mention sending reports to Feishu (飞书) — that would require invoking another (separate) Feishu/send-skill and credentials, which this skill does not request itself.
Install Mechanism
No install spec and no code files are present beyond metadata and documentation, so nothing is written to disk by an installer. This lowers risk and is consistent with being an instruction-only skill.
Credentials
The skill declares no required environment variables, credentials, or config paths. It does rely on agent-provided tools (memory_search, sessions_list/history) —appropriate for its purpose. There are no unrelated credential requests.
Persistence & Privilege
always is false and the skill does not request permanent presence or modify other skills/configs. It stores a lightweight thinking state conceptually, but as an instruction-only skill this is just guidance for agents and does not demand elevated platform privileges.
Assessment
This skill is an instruction-only meta-skill that will ask the agent to read its own memories, session history, and available skill list to form plans — that behavior is expected for a planning tool but can surface private information from your agent's memory and conversation history. Before installing: 1) confirm you are comfortable with other installed skills and stored memories being consulted; 2) remember the skill itself does not send data externally, but examples (e.g., posting to Feishu) would require calling other skills/services that may need credentials — review and limit those integrations; 3) note the package.json references a src/index.ts that isn't included, which is acceptable for an instruction-only skill but worth being aware of. If you want stricter limits, restrict the agent's permission to access memory/session history or avoid granting downstream skills (e.g., messaging integrations) access to sensitive destinations.Like a lobster shell, security has layers — review code before you run it.
latestvk978v650kznjkyga6qk3y9yn958455dj
License
MIT-0
Free to use, modify, and redistribute. No attribution required.