skills/letta-ai/skills/ai-news/Gen Agent Trust Hub

ai-news

Pass

Audited by Gen Agent Trust Hub on Apr 3, 2026

Risk Level: SAFEPROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [PROMPT_INJECTION]: The skill ingests untrusted external content (RSS feed items and video transcripts) to be summarized by the agent, creating a surface for indirect prompt injection.\n
  • Ingestion points: scripts/fetch-news.ts (RSS/Atom feeds) and scripts/fetch-transcript.ts (YouTube transcripts).\n
  • Boundary markers: None are explicitly defined in the instructions to separate external content from agent instructions.\n
  • Capability inventory: The scripts are restricted to network GET requests and local stdout; no dangerous system capabilities like file-write or shell-execution are exposed to the processed data.\n
  • Sanitization: Includes basic HTML stripping and character entity decoding on fetched content.\n- [EXTERNAL_DOWNLOADS]: Fetches AI news updates and YouTube metadata from well-known technology organizations and reputable media outlets including Hugging Face, OpenAI, Anthropic, Google DeepMind, VentureBeat, and The Verge.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 3, 2026, 04:05 AM