🚀 Proactive Agent Skill — 主动智能体
v1.0.0将 AI 智能体从被动任务执行者升级为可预测需求、持续自我进化的主动伙伴。内置 WAL 协议、Working Buffer、Autonomous Loop 等机制，让智能体在后台持续学习、提前规划并主动提供解决方案，显著降低人工干预频率。
9· 1.5万·197 当前·205 累计
by @fangkelvin下载技能包
最后更新
2026/4/21
安全扫描
OpenClaw
可疑
high confidenceNULL
评估建议
This skill's design is coherent for building a proactive agent, but it omits important operational details and security safeguards. Before installing or enabling it: 1) Confirm whether your environment has the 'openclaw' CLI and whether you trust it; 2) Expect the skill to create files (SESSION-STATE.md, MEMORY.md, working-buffer.md) and to create cron jobs — review and sandbox those actions first; 3) Do not provide email, calendar, or gateway credentials until the author documents exactly how t...详细分析 ▾
ℹ 用途与能力
The name and description (proactive agent, WAL, heartbeats, autonomous crons) align with the instructions: it describes memory files, compaction, and scheduled tasks. That said, the SKILL.md assumes integration points (email, calendar, weather, OpenClaw gateway/CLI) and system crons that are not declared in the skill metadata — the core purpose is coherent, but required capabilities are underspecified.
⚠ 指令范围
The instructions tell an agent to check email inboxes, calendars, weather, 'OpenClaw gateway status', create/modify files under a workspace, and install cron jobs that run 'openclaw run --task ...'. None of these external integrations, authentication steps, or required binaries are documented in the skill metadata. The instructions therefore ask the agent to access and act on resources (mail, calendar, system cron, filesystem, gateway) beyond what the registry claims.
✓ 安装机制
This is an instruction-only skill with no install spec or code files, which minimizes direct supply-chain risk. However, the SKILL.md implicitly depends on an 'openclaw' CLI and a runtime able to create cron entries and filesystem artifacts; those implicit dependencies are not declared.
⚠ 凭证需求
The skill declares no required environment variables or credentials, yet the instructions expect access to email, calendar, weather APIs, and the 'OpenClaw gateway' (all of which normally require tokens/credentials). This mismatch could lead to the agent requesting or using sensitive credentials at runtime without prior disclosure or scoping.
ℹ 持久化与权限
always:false (not force-included) and autonomous invocation is allowed (platform default). The skill's recommended behavior (creating cron jobs and persistent memory files like MEMORY.md and SESSION-STATE.md) would produce long-lived artifacts and scheduled autonomous actions on the host. While not an explicit 'always' privilege, these patterns create persistent side-effects that deserve user oversight.
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
latestv1.0.02026/3/2
NULL
● 可疑
安装命令
点击复制官方npx clawhub@latest install proactive-agent-skill
镜像加速npx clawhub@latest install proactive-agent-skill --registry https://cn.longxiaskill.com