MomClaw 元认知造虾师
v1.0.2用元认知引导发现值得被做成小龙虾的机会点，并将其收敛为可开箱即用的基准 Agent 小龙虾。
⭐ 0· 114·0 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
Name/description (metacognitive agent design / blueprinting) match the declared footprint: no binaries, no installs, no credentials. The templates and references are relevant to producing benchmark agent blueprints and handoff artifacts.
Instruction Scope
SKILL.md gives detailed runtime instructions that remain within the stated scope (questioning, scoring, architecture convergence, generating templates, and writing deliverables into a target Agent workspace). It explicitly instructs the agent to write files into the target workspace and to call governance/creator skills when proceeding to creation — expected for this purpose. Note: the skill will also主动 prompt the user about deployment platform and outline token/account preparation steps (it does not itself require credentials).
Install Mechanism
No install spec and no code files — instruction-only. No downloads or external packages, so there is no install risk.
Credentials
The skill declares no environment variables, credentials, or config paths. The content does guide users to prepare platform-specific tokens/accounts for deployment (Telegram/Feishu/Discord), but that is a user-driven step and not a required secret exposed by the skill itself.
Persistence & Privilege
always is false and model invocation is normal. The skill asks the agent to write documents into a target workspace (its own intended behavior) and to call governance/creator skills when appropriate; it does not request elevated platform-wide privileges or modify other skills' configs.
Assessment
This is an instruction-only planning skill that appears internally consistent. Before enabling or using it, verify: (1) the agent has explicit permission to write to the intended target workspace (so files are not accidentally written to the wrong project); (2) you understand that if you choose to proceed to deployment the workflow will ask you to prepare platform-specific tokens/accounts (Telegram/Feishu/Discord) — only provide those tokens to trusted tooling and follow least-privilege practices; (3) the downstream skills it references (openclaw-agent-governance and vertical-agent-creator) exist and are trusted in your environment, since the skill delegates creation to them; (4) review generated file paths/names before confirming writes to avoid overwriting important assets. Overall, the package is coherent and does not request unexplained secrets or install arbitrary code.Like a lobster shell, security has layers — review code before you run it.
agent-designvk97c0me4eqmq6ymafrmem2zb3x837ekwarchitecturevk97c0me4eqmq6ymafrmem2zb3x837ekwbenchmark-agentvk97c0me4eqmq6ymafrmem2zb3x837ekwlatestvk97c0me4eqmq6ymafrmem2zb3x837ekwmetacognitionvk97c0me4eqmq6ymafrmem2zb3x837ekwmomclawvk97c0me4eqmq6ymafrmem2zb3x837ekwplanningvk97c0me4eqmq6ymafrmem2zb3x837ekw
License
MIT-0
Free to use, modify, and redistribute. No attribution required.