Work Mode Switch
v2.0.1OpenClaw 智能体工作模式切换指南 - 12 种协作模式（自由/闲聊/讨论规划/研究/深度研究/写作/编码/翻译/执行/审核/自检/自学），支持灵活切换协作节奏
⭐ 0· 83·0 current·0 all-time
byJimmy@jijinbao-code
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 (work-mode switching guidance) matches the files and SKILL.md content: a documentation guide for 12 modes. There are no extraneous requirements (no env vars, binaries, or installs) that don't belong to this purpose.
Instruction Scope
SKILL.md is purely prescriptive documentation and does not instruct the agent to read credentials, call unexpected endpoints, or perform file I/O. Two minor points to note: (1) Deep Research mode describes 'independent推进直到交付' and multi-agent execution (i.e., proceed without frequent user interruptions), which implies autonomous behavior consistent with a research mode but could be surprising to some users; (2) some README lines and scenario examples use wording like '自动检�?/Auto-check' which slightly contradicts explicit statements in SKILL.md that file operations and system checks require user confirmation and that Self-Check only provides a checklist. These are wording/instruction consistency issues, not evidence of hidden actions.
Install Mechanism
Instruction-only skill with no install spec and no code files to execute. Package metadata and README are present but no installation scripts or remote downloads are defined, which is lowest risk.
Credentials
No environment variables, credentials, or config paths are requested. The SKILL.md references checking local files in a checklist but explicitly states it will not perform file operations without user confirmation; this is proportionate to a documentation skill.
Persistence & Privilege
Skill does not request always:true, does not modify agent/system configs, and contains no install-time hooks. It can be invoked by agents (default autonomous invocation), which is expected for a skill that describes autonomous research workflows — but the skill itself does not grant extra privileges.
Scan Findings in Context
[no_regex_findings] expected: Static regex scanner found nothing to analyze; this is expected because the skill is instruction-only and contains no executable code files.
Assessment
This skill is documentation-only and appears coherent with its stated purpose. Things to consider before installing: 1) Source trust — the package lists an owner ID and references 'SuperMike / OpenClaw Community' but has no homepage; if you require provenance, verify the publisher on ClawHub or the upstream repo before trusting. 2) Autonomy expectations — Deep Research mode and some README examples imply the assistant may run multi-step, low-interruption workflows; if you don't want autonomous long-running actions, keep agent autonomous-invocation disabled or confirm prompts before heavy operations. 3) File/system checks — Self-Check gives a checklist (e.g., MEMORY.md, HEARTBEAT.md) but SKILL.md says it will not perform file I/O without your consent; if you see an agent attempting to read local files, deny or require explicit confirmation. Overall the skill is low-risk, but verify publisher and be mindful of agent autonomy settings if you enable multi-stage/automatic behaviors.Like a lobster shell, security has layers — review code before you run it.
latestvk97bqb3r7eqyps85yebh1xkkc583zybp
License
MIT-0
Free to use, modify, and redistribute. No attribution required.