Universal Task OS

PassAudited by ClawScan on May 16, 2026.

Overview

This is an instruction-only task-planning framework that appears benign, with the main caution that it may guide a tool-enabled agent to use external tools or web search for broad user tasks.

This skill appears safe to install as a planning/methodology aid. Before using it with tool-enabled agents, review any proposed web searches, file operations, account actions, code changes, or uploads, especially when using private samples or working in regulated areas like finance, law, or healthcare.

Findings (1)

Artifact-based informational review of SKILL.md, metadata, install specs, static scan signals, and capability signals. ClawScan does not execute the skill or run runtime probes.

What this means

If the host agent has powerful tools, this framework could lead it to propose or perform tool-based steps across many task types.

Why it was flagged

The skill explicitly lets the agent choose tools or other skills when a pipeline unit needs external data or operations. This is broad, but it is disclosed, central to the workflow framework, and not paired with code, credentials, hidden endpoints, or automatic privileged execution.

Skill content
| 调用工具 | 需要外部数据或操作 | ... | 加载技能 | 领域专业工作流 |
Recommendation

Keep account changes, file edits, uploads, purchases, deployments, or other high-impact operations behind explicit user confirmation and normal tool permission checks.