scrapling-skill
Pass
Audited by Gen Agent Trust Hub on Oct 23, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill extracts content from external websites, which could potentially contain malicious instructions meant to subvert the agent's logic. . Ingestion points: External webpage content fetched via the scrapling CLI tool. . Boundary markers: No explicit delimiters are used to wrap extracted content in the instructions. . Capability inventory: Execution of local scripts and CLI tools (subprocess.run), and temporary file creation. . Sanitization: No content sanitization is performed on the extracted data.
- [COMMAND_EXECUTION]: The skill includes a Python script that executes shell commands for environment diagnostics. . Evidence: scripts/diagnose_scrapling.py uses the subprocess module to verify the scrapling installation and run smoke tests.
- [EXTERNAL_DOWNLOADS]: The skill instructions guide the user to install external software and browser runtimes. . Evidence: SKILL.md and references/troubleshooting.md provide commands to install the scrapling package and its browser dependencies using uv and the scrapling install command.
Audit Metadata