baoyu-url-to-markdown

Pass

Audited by Gen Agent Trust Hub on Mar 12, 2026

Risk Level: SAFECOMMAND_EXECUTIONREMOTE_CODE_EXECUTIONEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill uses child_process.spawn to launch browser binaries such as Chrome, Chromium, or Edge. It includes platform-specific logic in scripts/cdp.ts to locate these executables on macOS, Windows, and Linux.\n- [REMOTE_CODE_EXECUTION]: JavaScript is dynamically executed within the browser context via the Chrome DevTools Protocol (CDP) using the Runtime.evaluate command. This is used to run a content extraction script (cleanupAndExtractScript) on the target webpage.\n- [EXTERNAL_DOWNLOADS]: The skill's core purpose is to download and process content from arbitrary external URLs. Additionally, the documentation suggests using npx -y bun, which may download the Bun runtime environment if it is not already installed.\n- [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it retrieves untrusted data from the web and integrates it into the agent's context. Malicious content on a webpage could attempt to influence the agent's subsequent behavior.\n
  • Ingestion points: Arbitrary external URLs processed in scripts/main.ts.\n
  • Boundary markers: The output uses YAML frontmatter and markdown structure, but lacks explicit instructions for the agent to ignore instructions embedded in the webpage content.\n
  • Capability inventory: The skill has capabilities for file system access (mkdir, writeFile) and browser automation (spawn, CDP) via scripts/main.ts and scripts/cdp.ts.\n
  • Sanitization: scripts/html-to-markdown.ts performs HTML sanitization (removing scripts, styles, etc.) before conversion, but the final text content is not validated against prompt injection patterns.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 12, 2026, 07:52 AM