The Agent Skills Directory

[PROMPT_INJECTION]: The skill processes untrusted third-party data from X (tweets, threads, articles), creating an attack surface for indirect prompt injection.\n
Ingestion points: Reads content from x.com via scripts/fetch-x.ts (SKILL.md).\n
Capability inventory: Utilizes bun for script execution and jq for data processing.\n
Sanitization: Implements sanitize.ts to strip terminal escape sequences (OSC/CSI) and warns that fetched content must be treated as data, not instructions.\n- [EXTERNAL_DOWNLOADS]: Retrieves content and configuration from well-known Twitter/X service infrastructure.\n
Evidence: Accesses the X GraphQL API and retrieves JavaScript bundles from the official CDN (abs.twimg.com) to dynamically refresh operation IDs.

mktg-x