Web Fetcher
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.
A private, tokenized, intranet, or otherwise sensitive URL could be exposed to those fetch services if the user asks the skill to process it.
The helper sends the user-provided target URL through third-party URL-to-Markdown services. This is disclosed and purpose-aligned, but it creates an external data boundary.
METHODS = [
('r.jina.ai', ...),
('markdown.new', ...),
('defuddle', ...),
]Use the direct fetch flow for public pages; avoid submitting sensitive URLs unless you are comfortable with those external services handling them.
If you attach an authenticated or sensitive tab, the agent may read and summarize information visible in that tab.
The browser fallback can read content from a live browser tab, which may include content available because of the user's logged-in browser session.
If the user is using the Chrome relay/extension, ask them to attach the tab and then inspect the live rendered page. Snapshot the page and extract only the needed fields.
Attach only the specific tab you intend the agent to inspect, and avoid sensitive account pages unless that is explicitly your goal.
The helper may not run in an environment without Python, even though the metadata does not declare that prerequisite.
The skill documents running a bundled Python helper, while the registry requirements declare no required binaries. This is a minor setup metadata gap rather than hidden code installation.
python {baseDir}/scripts/fetch_url.py "https://example.com/article"Ensure Python is available before using the bundled script, or rely on other available browser/search tools.