Trending Skills

AdvisoryAudited by Static analysis on Apr 30, 2026.

Overview

No suspicious patterns detected.

Findings (0)

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

Installing the skill may require adding Python packages, a browser runtime, and system libraries.

Why it was flagged

The skill asks for unpinned package installation and a Chromium/system dependency install. This is expected for Playwright-based scraping, but users should understand it modifies the local Python/browser environment.

Skill content
pip install playwright
playwright install chromium --with-deps
Recommendation

Install in a virtual environment or other contained workspace, and review/pin dependency versions if reproducibility matters.

What this means

If the visited page or browser were compromised, reduced sandboxing could increase local exposure compared with default browser isolation.

Why it was flagged

The scraper launches Chromium with sandboxing disabled and an automation-hiding flag while visiting an external site. The browser use is purpose-aligned, but the launch options reduce containment.

Skill content
'--no-sandbox',
'--disable-setuid-sandbox',
'--disable-blink-features=AutomationControlled'
Recommendation

Prefer default sandboxing when possible, keep Playwright/Chromium updated, and run the skill in a low-privilege or contained environment.

NoteHigh Confidence
ASI01: Agent Goal Hijack
What this means

A malicious or misleading skill description on the source site could try to influence the agent or user if treated as trusted instructions.

Why it was flagged

The skill formats fetched skill descriptions and rule text for output/analysis. That remote content could contain instruction-like text, so it should remain informational.

Skill content
lines.append(detail.get("when_to_use"))
...
lines.append(f"  - {rule.get('file')}: {rule.get('desc')}")
Recommendation

Treat fetched descriptions and rules as untrusted web content; summarize or quote them without following any embedded instructions.

What this means

This does not show harmful behavior, but it slightly reduces confidence in the package’s provenance and maintenance hygiene.

Why it was flagged

The package header describes an unrelated AI Daily/smol.ai/Claude project rather than the Trending Skills functionality, suggesting copied or stale project metadata.

Skill content
AI Daily - AI资讯日报自动生成器
自动从 smol.ai 获取 AI 资讯,通过 Claude 分析分类,生成精美 HTML 页面
Recommendation

The publisher should update stale metadata; users who require strong provenance should verify the source before relying on it.