Back to skill
Skillv1.0.0
ClawScan security
Playwright Cli Skills · ClawHub's context-aware review of the artifact, metadata, and declared behavior.
Scanner verdict
BenignApr 14, 2026, 10:27 AM
- Verdict
- Benign
- Confidence
- medium
- Model
- gpt-5-mini
- Summary
- The skill's files and runtime instructions are consistent with a Playwright CLI automation/testing helper (including an Amazon scraping helper); it performs local browser automation and writes files to the user's home directory but does not request unrelated credentials or download remote code.
- Guidance
- This skill appears to do what it says: automate Playwright-driven browser interactions and includes an Amazon scraping helper. Before installing or running it, consider: 1) Playwright CLI and Node must already be installed on your machine. 2) The amazon-search scripts load a local storage-state/auth file (e.g., ~/amazon-auth.json or a project-specific path) and will use cookies/storage from it — those files can contain session tokens; only use auth state you trust and remove it after use. 3) The skill and its docs expose run-code and routing features that can execute arbitrary JS in pages and capture network/request bodies and cookies (use isolated browser sessions or a throwaway profile if you are concerned). 4) The included scripts run local shell commands and write CSV/traces to your home directory; review the JS and shell files before running them and run in a sandboxed account or VM if you need extra safety. If you want, I can point out the exact lines that load auth files, execute shell commands, and write output so you can review them quickly.
Review Dimensions
- Purpose & Capability
- okName/description (Playwright CLI automation & testing) match the provided SKILL.md, reference docs, and example scripts. The included amazon-search scripts use playwright-cli to open pages, load storage state, extract data, and write CSV output — all expected for a browser automation scraping helper.
- Instruction Scope
- noteThe SKILL.md and reference docs allow and document powerful operations: run-code (arbitrary JS in the page), loading/saving storage state, cookie/localStorage manipulation, recording traces (which include network bodies), and routing/removing headers (including cookie/authorization). Those are legitimate and expected capabilities for a Playwright/testing skill, but they also let the skill access and persist sensitive data (cookies, tokens, page content) and execute arbitrary page JS—so operators should understand those capabilities before use.
- Install Mechanism
- okNo install spec is present (instruction-only). The skill does not pull code from remote URLs or install packages itself. Risk is limited to running local binaries (playwright-cli, node) which the user must already install.
- Credentials
- noteThe skill declares no required env vars or credentials. Scripts use HOME and optional DEBUG env var to locate/read/write files (e.g., an auth state file and CSV output under the user's home). The Amazon scripts expect a saved auth state file; this is coherent but means the skill will access local storage-state files containing cookies/tokens if present.
- Persistence & Privilege
- okalways is false and the skill does not request to modify other skills or system-wide agent settings. It runs commands that create files under the user's home (auth state, CSV output, traces) which is normal for automation scripts.