Security audit

scrapling

Security checks for vulnerabilities and agentic risk

Overview

This is a real Scrapling documentation skill, but it warrants Review because it guides AI-assisted scraping and anti-bot bypass with weak authorization, privacy, and user-control boundaries.

Install only if you intend to scrape sites you own or are clearly authorized to access. Review every pip, Docker, browser-driver, MCP, proxy, and scraping command before execution; avoid anti-bot bypass and robots.txt overrides unless you have explicit permission. Do not use auto_save, cache, sessions, or MCP on private, authenticated, personal, regulated, or proprietary pages without a retention and access-control plan.

SkillSpector

Vulnerability Patterns

Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code

Findings (6)

Missing User Warnings

Medium

Confidence: 95% confidence
Finding: The skill explicitly promotes Cloudflare/anti-bot bypass, proxy rotation, large-scale scraping, and MCP-based AI control without any guardrails about authorization, legal constraints, privacy, or operational impact. In an AI-agent context, this materially increases the chance of unauthorized collection, evasion of access controls, and automated abuse of third-party systems.

Missing User Warnings

Medium

Confidence: 92% confidence
Finding: The installation guide explicitly labels `pip install scrapling` as something an AI can auto-execute, but it does not warn that this performs network access and modifies the local Python environment. In an agentic context, this increases the chance that a user or AI assistant will run package installation commands without informed consent, potentially changing the system state or pulling unreviewed code from package registries.

Missing User Warnings

Medium

Confidence: 95% confidence
Finding: The MCP setup instructions enable direct AI invocation of the Scrapling tool, including adding a user-scoped MCP server and launching it, but they do not warn that this grants AI clients a new capability to execute scraping operations and interact with browser/network resources. In an AI-integrated environment, silently expanding tool access can materially increase attack surface and lead to unintended automated actions.

Missing User Warnings

Medium

Confidence: 92% confidence
Finding: The guide explicitly encourages `auto_save=True` to persist page element snapshots into a local database, but it does not warn users that scraped page content or structural metadata may be stored on disk. In a scraping framework that targets protected or dynamic sites, this can lead to unintentional retention of sensitive, proprietary, or regulated data and increase local exposure if the database is later accessed or exfiltrated.

Missing User Warnings

Medium

Confidence: 90% confidence
Finding: The MCP section explicitly instructs users to connect Scrapling to Claude/Cursor and send extracted website data into AI workflows, but it does not warn that scraped content may contain sensitive, proprietary, or personal data that will be transmitted to external AI tooling. In a scraping framework, this omission is security-relevant because users may unknowingly exfiltrate collected data outside their local environment.

Missing User Warnings

Medium

Confidence: 90% confidence
Finding: The troubleshooting guidance explicitly recommends setting `robots_txt_obey = False` to improve crawl speed, and only adds a weak qualifier without explaining legal, contractual, or access-control implications. In a scraping framework that also advertises anti-bot bypass and large-scale crawling, this materially encourages behavior that may violate site policies and increase the chance of unauthorized or abusive access.

Static analysis

No suspicious patterns detected.