ai-news-briefing

Fail

Audited by Gen Agent Trust Hub on Feb 16, 2026

Risk Level: HIGHPROMPT_INJECTIONREMOTE_CODE_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
  • PROMPT_INJECTION (HIGH): The skill is highly susceptible to Indirect Prompt Injection.\n
  • Ingestion points: Fetches content from 17+ external RSS feeds in collectors/rss_common.py.\n
  • Boundary markers: Uses only a simple [AI_NEWS_CANDIDATES] header in pipeline/prompt_builder.py without instructions for the agent to ignore commands within the external data.\n
  • Capability inventory: The resulting ai_input is directly fed into the AI prompt. Maliciously crafted news items could hijack the agent's behavior if it has write or execution capabilities.\n
  • Sanitization: Only basic HTML tag stripping is performed; no validation or filtering of natural language instructions is present.\n- REMOTE_CODE_EXECUTION (MEDIUM): Insecure XML parsing in collectors/rss_common.py.\n
  • Evidence: Uses xml.etree.ElementTree.fromstring to process external RSS feeds. This module is not secure against maliciously constructed data and is vulnerable to XML-based attacks like Denial of Service (Billion Laughs).\n- EXTERNAL_DOWNLOADS (MEDIUM): Recommends installation from an untrusted GitHub user.\n
  • Evidence: The README.md suggests installation via npx skills add vc999999999/ai-news-skill. The user vc999999999 is not a verified or trusted entity within the security policy.
Recommendations
  • AI detected serious security threats
Audit Metadata
Risk Level
HIGH
Analyzed
Feb 16, 2026, 01:05 PM