Back to skill
Skillv1.0.0

ClawScan security

Lightpanda Browser · ClawHub's context-aware review of the artifact, metadata, and declared behavior.

Scanner verdict

SuspiciousApr 18, 2026, 2:38 AM
Verdict
suspicious
Confidence
medium
Model
gpt-5-mini
Summary
The skill's instructions are coherent with a headless-browser tool, but provenance and install practices (nightly binary download, no checksums or homepage) are under-specified and increase risk before you execute the supplied binary or pull the Docker image.
Guidance
This SKILL.md is plausible for a headless scraper, but verify provenance before running anything: 1) Ask the publisher for an official homepage/repository and cryptographic checksums or signatures for the release binaries (the SKILL.md points at a nightly GitHub release but provides no verification). 2) Prefer a stable, signed release over 'nightly'. 3) Inspect the Docker image source (Dockerfile/repo) before pulling, and run unknown binaries in an isolated VM or container with restricted network and filesystem access. 4) If you must try it locally, scan the binary (antivirus/sandbox) and avoid running it as root. 5) If you need higher assurance, request source code or reproducible build instructions so you or a reviewer can audit the implementation.

Review Dimensions

Purpose & Capability
okName, description and runtime instructions align: SKILL.md documents downloading a standalone binary and shows commands and Python examples that call that binary for web scraping and content extraction. No unrelated credentials or system accesses are requested.
Instruction Scope
noteInstructions only direct downloading a binary from a GitHub releases URL, running the binary, and optionally running a Docker image. They do not request unrelated files, env vars, or sensitive data. However the install instructions run an unverified 'nightly' binary by curl+chmod with no checksum/signature verification—this broadens what gets executed on the host.
Install Mechanism
concernNo formal install spec in metadata; SKILL.md tells users to curl a 'nightly' release binary and to pull a Docker image 'lightpanda/browser:nightly'. While GitHub releases and Docker Hub are common distribution points, using a nightly binary without checksum/signature or an authoritative homepage/source is higher risk and makes supply-chain/execution trust unclear.
Credentials
okThe skill declares no required env vars, credentials, or config paths and the instructions do not read secrets. This is proportionate for a local scraping tool.
Persistence & Privilege
okSkill is not marked always:true and does not request persistent system-wide changes. It simply instructs to place a binary in the current directory or run a Docker container—normal for this purpose.