web-to-markdown

Pass

Audited by Gen Agent Trust Hub on Sep 15, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTIONDYNAMIC_EXECUTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill processes content from external websites which could contain malicious instructions designed to influence the agent's behavior.\n
  • Ingestion points: Converted Markdown output from external URLs (SKILL.md, Workflow section).\n
  • Boundary markers: The skill instructions do not specify any delimiters or warnings to the agent to ignore instructions embedded in the processed web content.\n
  • Capability inventory: The skill can execute shell commands like mkdir, web2md, ls, wc, and npm (SKILL.md, Workflow section).\n
  • Sanitization: While the skill uses libraries to clean formatting (Readability, Turndown), there is no mention of filtering or sanitizing text content for potential prompt injections.\n- [COMMAND_EXECUTION]: The skill uses the shell to run the web2md CLI, manage directories (mkdir), and verify output (ls, wc). These are necessary for the skill's primary function (SKILL.md, Workflow section).\n- [DYNAMIC_EXECUTION]: The skill relies on Puppeteer to render JavaScript from external websites, which involves executing remote code in a browser context as part of the content extraction process (README.md, Technical Details).
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 15, 2026, 12:54 AM
Security Audit — agent-trust-hub — web-to-markdown