critical-writing
PassAudited by ClawScan on May 8, 2026.
Overview
This instruction-only writing skill is coherent and does not request system access, though it may force a multi-step critique workflow for broad writing requests.
This appears safe as an instruction-only writing workflow. Before installing, consider whether you want an agent to use a mandatory multi-step questioning, drafting, debate, and rewrite process for many writing requests; if you usually want fast direct drafts, this skill may feel overly controlling but not security-dangerous.
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.
The agent may ask extra questions and run the debate-and-rewrite process instead of immediately producing a draft.
These instructions make the skill apply to a broad range of writing tasks and require its workflow even when the user asks for a direct draft. This steers response style but remains consistent with the stated writing-workflow purpose.
触发场景（必须使用本 skill）... 用户需要写文章、报告、分析、方案、帖子、邮件等任何正式或半正式内容 ... 不要跳过任何步骤，即使用户说"直接写就好"，也必须完成前置三问再动笔。
Use this skill when you want structured, iterative critique; avoid or disable it for quick writing tasks where you do not want the mandatory workflow.