v1.0.2

亲笔写的

ReviewClawScan verdict for this skill. Analyzed May 1, 2026, 5:06 PM.

Analysis

This is an instruction-only writing skill with no technical permissions, but it explicitly helps conceal AI authorship and its examples add invented-looking personal details, so users should review it carefully before use.

GuidanceBefore installing, decide whether your intended use is compatible with your employer, school, or platform rules on AI-generated content. If you use it, keep the no-fabrication rule explicit: do not let it invent personal stories, data, citations, or experiences, and disclose AI assistance where required.

Findings (2)

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.

Abnormal behavior control

Checks for instructions or behavior that redirect the agent, misuse tools, execute unexpected code, cascade across systems, exploit user trust, or continue outside the intended task.

Human-Agent Trust Exploitation

SeverityMediumConfidenceHighStatusConcern

SKILL.md

> 让领导看不出是AI写的。

The skill explicitly presents its purpose as making AI-written text undetectable to a leader, which can help users misrepresent authorship rather than merely improve writing quality.

User impactUsing the skill as written could violate workplace, school, or platform expectations around AI disclosure and could mislead readers about who wrote the text.

RecommendationUse only for legitimate editing; require truthful disclosure where applicable and ask the agent not to help conceal AI use when policy or context requires transparency.

Human-Agent Trust Exploitation

SeverityMediumConfidenceHighStatusConcern

examples/before_after.md

改后（知乎风格）: "我自己的经验——我在一家MCN做内容策划..." ... "McKinsey 2024年的报告说得很清楚..."

The example rewrites generic source text into specific first-person experience and named-report claims. This conflicts with the SKILL.md no-fabrication instruction and may teach the agent to add unsupported facts to appear more human.

User impactThe agent may produce polished text that includes personal anecdotes, sources, or details the user did not provide, causing readers to trust claims that may be false.

RecommendationInstruct the agent to mark missing details as placeholders, avoid inventing personal experiences or citations, and keep all factual claims traceable to user-provided information.