webcrawler

Pass

Audited by Gen Agent Trust Hub on Apr 29, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: The skill presents an indirect prompt injection surface as it processes untrusted data from external websites.
  • Ingestion points: External content is fetched in scripts/crawl_docs.py and scripts/extract_page.py using the requests library.
  • Boundary markers: The skill wraps harvested content with YAML frontmatter delimiters (---) in the generated output files.
  • Capability inventory: The crawler scripts (scripts/crawl_docs.py, scripts/extract_page.py) have network access for fetching URLs and file system write capabilities for saving documentation.
  • Sanitization: The scripts use BeautifulSoup and html2text to remove HTML tags and scripts, but the remaining natural language content is not filtered for malicious instructions targeting the AI agent.
  • [EXTERNAL_DOWNLOADS]: The skill performs HTTP GET requests to retrieve documentation from arbitrary, user-supplied URLs. This is the primary function of the agent for content harvesting.
  • [COMMAND_EXECUTION]: The skill instructions involve executing provided Python scripts to automate crawling, extraction, and post-crawl filtering tasks.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 29, 2026, 09:44 PM
Security Audit — agent-trust-hub — webcrawler