Clip to Your Vault
Security checks across malware telemetry and agentic risk
Overview
The clipper’s main purpose is coherent, but optional Douyin and browser-rendering features rely on unreviewed external tools and session/browser access that are not tightly scoped.
Install the core clipper only if you are comfortable with it writing to your chosen vault folder and sending selected URLs to the documented fetch services. Treat the Douyin downloader and CDP browser features as higher risk: inspect and pin external tools, use isolated accounts/browser profiles, and avoid enabling cookie or remote-debugging workflows unless you need them.
VirusTotal
VirusTotal findings are pending for this skill version.
Risk analysis
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.
Clipping a URL can create local Markdown files and download media into the configured vault path.
The skill is expected to fetch remote content and write notes/media locally, which is purpose-aligned for a web clipper but gives it file-write and download authority.
Download all images and videos to `ATTACHMENTS/` ... Write to `X_DIR/{title}.md`Use a dedicated clippings directory, review overwrite prompts, and avoid clipping sensitive/private URLs unless you are comfortable with the configured fetch methods.
A single clipping request may create extra GitHub notes and links beyond the original URL.
A clipped page can trigger additional handler work and persistent notes based on links found in the page content.
If the content contains a `github.com/{owner}/{repo}` link, auto-trigger the GitHub Handler to create a GitHub note, then add bidirectional wikilinksReview auto-created linked notes, and consider requiring confirmation before clipping linked repositories if the page content is untrusted.
Account session cookies can let tooling act as the logged-in user or expose sensitive account/session data if mishandled.
The optional Douyin workflow asks the user to collect/use cookies through an external tool, but the artifacts do not define which cookies are accessed, how they are stored, or what account scope is used.
# 3. Get cookies ... pip install playwright && python -m playwright install chromium ... python -m tools.cookie_fetcher --config config.yml
Only enable this if necessary, inspect the downloader, use a separate low-risk account/browser profile, and avoid sharing or committing cookie/config files.
A changed or compromised dependency could run code locally or mishandle downloaded content/session data.
The optional Douyin setup executes unpinned third-party code and dependencies outside the reviewed artifact set, and that tool is then used in the clipping workflow.
git clone https://github.com/jiji262/douyin-downloader.git ~/tools/douyin-downloader ... pip install -r requirements.txt
Pin to a trusted commit, review the third-party tool and requirements, install in an isolated environment, and avoid running it with unnecessary privileges.
If enabled against a normal browser profile, the clipping workflow could interact with or read pages using existing browser sessions.
The optional CDP path creates a browser-control channel, but the artifacts do not specify an isolated Chrome profile, page limits, or protections for logged-in browser state.
web: cdp_enabled: true; cdp_url: "http://localhost:3456" ... Requires Chrome with remote debugging enabled.
Keep CDP disabled unless needed, use a separate browser profile with no sensitive logins, bind only to localhost, and close the debugging browser after use.