scraperapi-agent-onboarding

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (high risk: 0.85). The required runtime workflow for Path A (MCP Server) hands the agent a tool that fetches arbitrary user-supplied URLs and returns scraped HTML/markdown into the agent’s tool results, which the agent then places into its LLM context—i.e., outsider-authored free text from public web pages via the MCP scrape/structured-data tools.

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (high risk: 0.90). The skill invokes a remote MCP server at https://mcp.scraperapi.com/mcp (used with "npx mcp-remote ...") which is contacted at runtime to supply tool definitions/behavior that can directly control agent prompts/behavior, and the local alternative requires fetching and executing remote code via "pip install scraperapi-mcp-server" then "python -m scraperapi_mcp_server", so this is a runtime external dependency that controls the agent.