安全扫描
OpenClaw
安全
high confidenceThis is an instruction-only coaching 技能 whose required resources and behavior are consistent with its 状态d purpose and it does not 请求 系统 访问 or 凭证s.
评估建议
该技能看起来逻辑清晰且技术风险低,因为它仅提供指令,无需凭据或安装。启用前,请决定是否允许智能体自主启动夜间审计,并使用强硬/强制语气:若不允许,请限制其自主调用,或在有限范围内测试。留意过多或不需要的提示,并确认你可禁用或调整触发条件(例如,何种“表现”会触发教练)。若想更谨慎,先在非关键智能体/配置中试用,观察频率与语气。...详细分析 ▾
✓ 用途与能力
Name/description clAIm a behavioral coaching function and the 技能 is instruction-only with no binaries, env vars, or 安装s — all of which is proportionate to a conversational coach that enforces action.
ℹ 指令范围
技能.md gives concrete 运行time behavior (force an Action Item, timebox tasks, restrict options). It also says the 技能 may '主动发起' a nightly 审计 if the user has not 报告ed — this grants the 代理 discretion to proactively message the user. The instructions do not ask the 代理 to read files, use 凭证s, or transmit data externally, but they do permit persistent/unsolicited prompts and strong/firm language; consider whether you want that level of proactivity/tone.
✓ 安装机制
No 安装 spec and no code files — lowest-risk delivery 模型 (instruction-only).
✓ 凭证需求
无需环境变量、凭据或配置文件路径——恰如其分地扮演一位纯粹的对话教练。
ℹ 持久化与权限
always:false(非强制安装)。该技能允许自主调用(disable-model-invocation:false),且指令明确提及启动夜间审计;自主调用对技能而言属正常,但结合主动审计意味着它可发起未经请求的互动——如需限制,请核查代理的调用策略。
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
latestv1.0.02026/4/24
初始发布:执行力唤醒教练,帮助截断分析瘫痪,强制闭环。
● 无害
安装命令
点击复制官方npx clawhub@latest install execution-coach
镜像加速npx clawhub@latest install execution-coach --registry https://cn.longxiaskill.com镜像同步中