baoyu-danger-x-to-markdown

Pass

Audited by Gen Agent Trust Hub on Jun 26, 2026

Risk Level: SAFECREDENTIALS_UNSAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
  • [CREDENTIALS_UNSAFE]: The skill utilizes a hardcoded public Bearer token in scripts/constants.ts to facilitate requests to the X.com API. This is a known public token used by the X web application for unauthenticated requests. Additionally, the skill handles sensitive session cookies (auth_token, ct0) for authentication, which are stored locally in the user's application data directory to maintain session state across uses.
  • [COMMAND_EXECUTION]: In scripts/paths.ts, the skill employs execSync to run cmd.exe and wslpath specifically for resolving path discrepancies when running in a Windows Subsystem for Linux (WSL) environment. These are static commands used for environment discovery and do not involve untrusted user input.
  • [EXTERNAL_DOWNLOADS]: The skill dynamically fetches and parses JavaScript bundles from Twitter's official CDN (abs.twimg.com) to extract necessary GraphQL query IDs and feature switches. It also downloads media assets from Twitter's media servers (pbs.twimg.com, video.twimg.com) to local storage when the --download-media flag is used. All remote requests are directed to well-known service domains associated with the skill's primary function.
  • [PROMPT_INJECTION]: The skill processes external, untrusted content from X.com, which presents an indirect prompt injection surface. This is an inherent risk in content-fetching tools. The skill mitigates this by requiring explicit user consent via a disclaimer before any operations are performed and by isolating the fetched content within standard Markdown formatting.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 26, 2026, 08:18 AM
Security Audit — agent-trust-hub — baoyu-danger-x-to-markdown