playwright-local

SUSPICIOUS. The core Playwright setup is coherent and mostly benign, using official package sources for the main tool, but the skill goes beyond ordinary automation into stealth/anti-bot evasion and includes third-party stealth dependencies plus an inconsistent MCP install example. No direct credential theft or exfiltration is shown, but the publisher/scope mismatch around optional packages and bypass guidance makes the overall skill medium risk rather than benign.

The provided fragment is strongly indicative of an abuse-enabling browser automation toolkit: it teaches evasion of anti-bot controls (stealth/timing/proxy rotation), demonstrates authenticated login using real credentials, persists harvested session cookies to disk, and includes CAPTCHA-solving integration via third-party services. While it does not clearly show conventional malware (e.g., backdoors/system damage) in the fragment itself, it materially enables credential/session misuse and policy-violating scraping at scale. Treat as high operational security risk for any supply-chain dependency and require strict review for legitimacy and usage constraints.

No clear indicators of classic malware (exfiltration, system compromise, or persistence) exist in the provided fragment. However, it is explicitly engineered to evade bot-detection by spoofing multiple high-signal browser fingerprint/probing surfaces at page load and by using headless automation and optional proxy routing (including credentials). It also recommends practices commonly used for abusive automation (residential proxies and CAPTCHA/anti-captcha). Additionally, disabling the browser sandbox via launch flags creates an operational security hardening concern. Overall, this module should be treated as high-abuse-risk software rather than malware, and its use should be tightly controlled and reviewed for compliance and safe runtime configuration.