Skills
Navigate
skills/muratcankoylan/baoyu-skills/baoyu-danger-x-to-markdown/Gen Agent Trust Hub

baoyu-danger-x-to-markdown

Pass

Audited by Gen Agent Trust Hub on Jun 28, 2026

Risk Level: SAFECOMMAND_EXECUTIONCREDENTIALS_UNSAFEEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: Spawns browser processes (Chrome, Edge, or Chromium) to capture authentication cookies via the Chrome DevTools Protocol (CDP). The executable path can be influenced via environment variables.
  • Evidence: scripts/cookies.ts uses child_process.spawn to launch the browser with remote debugging enabled.
  • [CREDENTIALS_UNSAFE]: Includes a hardcoded bearer token. While this is a public credential used by the official X web client, hardcoding any authentication token is a security anti-pattern.
  • Evidence: scripts/constants.ts defines DEFAULT_BEARER_TOKEN.
  • [EXTERNAL_DOWNLOADS]: Fetches and parses remote JavaScript bundles from abs.twimg.com at runtime to dynamically resolve API query IDs and feature flags.
  • Evidence: scripts/graphql.ts fetches content from X's content delivery network.
  • [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection as it processes untrusted data from an external platform (X) and converts it to markdown for agent or user consumption.
  • Ingestion points: Tweets and articles are fetched from X's GraphQL API in scripts/graphql.ts and scripts/thread.ts.
  • Boundary markers: The output is structured as markdown with YAML frontmatter delimiters (---).
  • Capability inventory: Access to subprocess spawning (scripts/cookies.ts), network requests via fetch (scripts/http.ts), and file system writes (scripts/main.ts).
  • Sanitization: Relies on JSON.stringify for metadata and simple markdown formatting for content body.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 28, 2026, 03:19 AM
Security Audit — agent-trust-hub — baoyu-danger-x-to-markdown