The Agent Skills Directory

[SAFE]: The skill provides documentation and tool definitions for the Crawlio MCP server. It does not contain any executable code, malicious logic, or hidden instructions.- [PROMPT_INJECTION]: The skill exhibits an attack surface for indirect prompt injection (Category 8) because it ingests untrusted data from external websites via tools like start_crawl, trigger_capture, and analyze_page. Evidence: 1. Ingestion points: Tools that fetch web content in SKILL.md; 2. Boundary markers: Absent; 3. Capability inventory: Network access (execute_api) and file system operations (export_site, extract_text_from_image); 4. Sanitization: Not specified. This is a functional requirement of a crawler and is considered safe within context.- [COMMAND_EXECUTION]: The extract_text_from_image and export_site tools utilize absolute file paths for reading and writing data. This capability is inherent to the skill's primary purpose of performing OCR on local images and exporting crawl results.

crawlio-mcp