Skillv1.0.0

ClawScan security

web-to-markdown · ClawHub's context-aware review of the artifact, metadata, and declared behavior.

Scanner verdict

ReviewMar 5, 2026, 1:30 PM

Verdict: Review
Confidence: medium
Model: gpt-5-mini
Summary: The skill mostly does what it says (scrape pages, extract/download images) but has mismatched dependency claims and uses third‑party fetching (r.jina.ai), which raises privacy and clarity concerns.
Guidance: This skill appears to implement the advertised scraping and image‑downloading features, but note two issues before installing: (1) the scripts require the third‑party Python package 'requests' (and scrape.py needs 'scrapling') despite a claim they only use the standard library — install those packages or the scripts will fail; (2) the tool sends page URLs/content to external conversion/fetch services (r.jina.ai, markdown.new, defuddle.md) — any content you fetch will be visible to those services, so avoid scraping sensitive/internal URLs. Also review and confirm the output directory (~/.openclaw/images) is acceptable, and be mindful of site terms and server load when bulk downloading.

Review Dimensions

Purpose & Capability: noteName/description match the included scripts and instructions: converting pages to Markdown and extracting/downloading images. However, SKILL.md incorrectly states that extract_images.py and download_images.py use only the Python standard library, while both scripts import the third‑party 'requests' package (and scrape.py requires 'scrapling'). This is an incoherence between claimed dependencies and actual requirements.
Instruction Scope: noteInstructions are narrowly scoped to web scraping, URL extraction, and downloading images. They explicitly instruct using external conversion services (markdown.new, defuddle.md, r.jina.ai). This means page URLs and page contents will be sent to those remote services — a privacy/ data‑sharing implication that is declared but should be highlighted to users. The scripts write files to a default local directory (~/.openclaw/images), which is expected for a downloader.
Install Mechanism: noteThere is no install spec (instruction-only), so nothing is auto-downloaded/installed by the platform. The included scripts, however, depend on third‑party Python packages: 'requests' (used by extract_images.py and download_images.py) and 'scrapling' (required by scrape.py). SKILL.md partly documents scrapling installation but incorrectly claims the first two scripts need only the standard library. This mismatch could cause runtime failures and surprises for users.
Credentials: okThe skill requests no environment variables or credentials and does not access system config paths beyond creating a per‑user output directory. There are no undeclared credential requirements.
Persistence & Privilege: okalways is false, the skill is user-invocable and does not request persistent/privileged presence or modify other skill configurations. It only writes files to its output directory.