Bright-Data-MCP-Claude-Skill-deep-research
ReviewAudited by ClawScan on May 10, 2026.
Overview
The skill is a coherent Bright Data web-scraping guide, but it includes anti-bot browser automation and logged-in/session-persistent scraping without clear boundaries.
Use this only if you intentionally want Bright Data MCP-powered web scraping. Set clear domains, URL counts, and output limits; avoid logged-in or private pages unless you explicitly approve them; and ensure scraping complies with website terms, robots.txt, and applicable law.
Findings (2)
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.
The agent could scrape or interact with websites in ways that trigger site blocks, violate site terms, or go beyond the user's intended passive research.
The skill directs the agent to perform broad scraping and interactive browser actions, including anti-bot use. Although this fits the scraping purpose, the artifacts do not require explicit approval, domain allow-lists, or stop conditions before interactive/anti-bot workflows.
"Research requiring anti-bot protection" and "Deep Mode (20-50 URLs, browser automation)" using "scraping_browser_click"
Use explicit URL/domain and result limits, require confirmation before deep/browser mode or clicks, and follow robots.txt, site terms, and rate limits.
If connected to a real account, the agent could access private or account-specific pages and leave session state available for later tool calls.
The workflow contemplates authenticated browsing and persistent cookies/session state, but the artifacts do not define which accounts may be used, how user consent is obtained, or how sessions are cleared.
"Content is behind login walls" ... "re-authenticate if needed" ... "Use Session Persistence: Browser keeps cookies/state across calls"
Avoid logged-in/private pages unless explicitly authorized, use a dedicated low-privilege account if needed, and clear browser sessions/cookies after the task.