Skill flagged — suspicious patterns detected
ClawHub Security flagged this skill as suspicious. Review the scan results before using.
Nonblocking Tasks
v1.3.0三层机制非阻塞任务管理器 — 主会话永远空闲 + 新消息立即响应 + 主动推送结果。适用于所有通道、所有用户的任务管理,确保主会话不被任何操作阻塞,每条消息都能被即时响应。
⭐ 0· 105·0 current·0 all-time
MIT-0
Download zip
LicenseMIT-0 · Free to use, modify, and redistribute. No attribution required.
Security Scan
OpenClaw
Suspicious
medium confidencePurpose & Capability
Most required actions (tracking active tasks, heartbeat checks, spawning background agents, and notifying users) match the stated purpose. However, the skill instructs reading user conversation logs from ~/.openclaw/memory-tdai (cross-channel recall) and creating a system cron 'nonblocking-task-monitor' and modifying AGENTS.md / MEMORY.md automatically. Those actions go beyond a simple task manager and are not clearly justified by the description; reading global conversation archives is especially privacy-sensitive.
Instruction Scope
SKILL.md instructs the agent to automatically create/update files in <workspace> (active_tasks.md, HEARTBEAT.md, AGENTS.md, MEMORY.md) and to create a cron job at install-time or on first message. It also instructs reading ~/.openclaw/memory-tdai/conversations/YYYY-MM-DD.jsonl (user conversation history across channels). The skill will auto-trigger configuration on the first user message. Automatic modification of workspace files, creation of system cron jobs, and reading home-directory conversation logs without explicit, declared consent is scope creep and privacy-sensitive.
Install Mechanism
Instruction-only skill with no install spec or code files—no third-party downloads or archive extraction. From an installation code-distribution perspective this is low-risk.
Credentials
The skill declares no required env vars or credentials but expects to send notifications to enterprise channels (飞书/企微) and to access ~/.openclaw conversation logs and create cron jobs. Those capabilities imply access to channel integrations and filesystem/system cron control that are not declared or justified. Requesting implicit access to cross-channel memory and system scheduling is disproportionate for an install-less instruction-only skill.
Persistence & Privilege
Although always:false, the skill instructs creation of a persistent cron job running every 2 minutes and automatic modification of AGENTS.md / MEMORY.md and active_tasks.md. Creating system cron entries and a persistent monitor increases blast radius and requires elevated privileges; this persistent presence combined with cross-channel memory reads is a notable privilege escalation relative to a simple skill.
What to consider before installing
Before installing, consider these actions:
- Do not install blindly. The skill will auto-create files in your workspace and add a system cron job; ask for an explicit opt-in and a dry-run mode.
- Confirm what 'spawn child agents' means in your OpenClaw environment and whether those agents get extra privileges or access to secrets.
- Review and backup AGENTS.md and MEMORY.md prior to installation; require the skill to open a PR or show diffs rather than auto-writing those files.
- Deny or audit access to ~/.openclaw memory files if you do not want cross-channel conversation recall; this is privacy-sensitive and not justified explicitly in the description.
- Verify how notifications to 飞书/企微 will be delivered and what credentials are needed; prefer explicit configuration of notification endpoints rather than implicit assumptions.
- If you want to try it, run in an isolated environment or test workspace without real user data and without granting permission to modify system cron, then inspect behavior and logs.
I have medium confidence in this assessment; additional artifacts that would raise or lower confidence: concrete implementation code showing how 'spawn', cron creation, and cross-channel reads are performed (would increase confidence), or an explicit, permissioned installation workflow that prompts the user before making changes (would lower concern).Like a lobster shell, security has layers — review code before you run it.
latestvk970z8vzjvhpfkezhpk5rf2ejn845s8r
License
MIT-0
Free to use, modify, and redistribute. No attribution required.
