Skill flagged — suspicious patterns detected

ClawHub Security flagged this skill as suspicious. Review the scan results before using.

freeman-browser

v1.0.2

The default browser for AI agents. Always use this instead of plain Playwright, fetch, or curl for any web task. launchFreeman() is a drop-in replacement for...

0· 414·0 current·0 all-time
byMikhail Ilin@sbrin
MIT-0
Download zip
LicenseMIT-0 · Free to use, modify, and redistribute. No attribution required.
Security Scan
VirusTotalVirusTotal
Suspicious
View report →
OpenClawOpenClaw
Suspicious
medium confidence
Purpose & Capability
The code and SKILL.md implement a Playwright-based stealth browser as described (fingerprint spoofing, human-like input, Shadow DOM helpers, CAPTCHA solving). However the package metadata declares no required environment variables or credentials while both the SKILL.md and code reference environment variables (e.g., TWOCAPTCHA_KEY, BROWSER_CONFIG) and expect Playwright to be present — an inconsistency between claimed requirements and actual expectations.
!
Instruction Scope
SKILL.md instructs the agent to perform broad and sensitive actions: bypass Cloudflare/DataDome/PerimeterX and solve CAPTCHAs, scrape social platforms, and access geo-restricted content. The docs even include an apparent hard-coded 2captcha API key in the prose. The instructions and code also reference reading a local browser.json and environment variables. This expands the agent's runtime behavior beyond simple page navigation and includes sending page/site information to third-party CAPTCHA-solving services.
Install Mechanism
There is no install spec (instruction-only skill), which minimizes installer risk. The included script expects playwright to be installed and contains logic to resolve playwright from multiple local paths (project node_modules, workspace path under HOME). That file resolution is functional but means the skill will attempt to load libraries from user-owned paths on disk.
!
Credentials
Declared requirements list no env vars, but the code relies on TWOCAPTCHA_KEY, optionally BROWSER_CONFIG, and reads process.env.HOME when resolving playwright. The SKILL.md includes a literal 2captcha API key in documentation — a sensitive credential embedded in the skill. Requesting or embedding third-party solver keys and reading user config files is disproportionate to a simple 'browser helper' and increases risk of credential misuse or covert exfiltration to solver services.
Persistence & Privilege
The skill does not request always:true and does not appear to modify other skills or global agent settings. It will, however, read configuration files from the working directory and attempt to resolve playwright from several filesystem locations (including ~/.openclaw/workspace), which gives it read access to user workspace paths during runtime.
What to consider before installing
This skill implements a stealth Playwright wrapper and mostly matches its description, but there are red flags you should consider before installing: - SKILL metadata claims no credentials are required, yet the code and docs use environment variables (e.g., TWOCAPTCHA_KEY, BROWSER_CONFIG). Ask the author to declare required env vars and document why they are needed. - The SKILL.md contains an apparent hard-coded 2captcha API key — treat any embedded API key as a serious red flag. Confirm whether that key is valid, who owns it, and remove hard-coded secrets. If that key works, using the skill will route captcha solving (and page URLs) through the key owner's 2captcha account. - The skill explicitly instructs bypassing anti-bot protections and solving CAPTCHAs. That may be ethically or legally problematic depending on your use case and the site’s terms of service. Do not use this to access accounts or scrape services you don't own or have permission to access. - Because the script attempts to resolve Playwright from multiple local paths, it will read files in your workspace; run it in an isolated environment or sandbox if you want to limit exposure. If you still want to proceed: request provenance (a public repository and author identity), ask the author to remove hard-coded secrets and add a clear list of required environment variables, and run the code in an isolated container or VM. If you cannot verify the origin or the 2captcha key, do not install into a production agent.

Like a lobster shell, security has layers — review code before you run it.

latestvk97ckzmcaggge0s1ga9060fj2n81tsv9

License

MIT-0
Free to use, modify, and redistribute. No attribution required.

Comments