prospective-news-coverage
Pass
Audited by Gen Agent Trust Hub on Oct 16, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSPROMPT_INJECTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill contains purely instructional markdown and documentation for political news analysis without any embedded scripts or executable code.
- [EXTERNAL_DOWNLOADS]: References official Swedish government (Riksdag, Regeringen) and international organization (EU, Nordic Council) websites as trusted sources for primary data.
- [EXTERNAL_DOWNLOADS]: Includes links to technical documentation for agentic workflows hosted on a well-known service (GitHub).
- [PROMPT_INJECTION]: Uses strong behavioral framing (e.g., the 'AI FIRST Quality Principle') to enforce multi-pass reasoning and analytical rigor, which are benign quality-control instructions rather than safety bypass attempts.
- [INDIRECT_PROMPT_INJECTION]: The skill instructions involve processing external data from political agendas and news sources, which represents a potential ingestion surface for indirect prompt injection, although the skill itself does not provide tools for automated execution of such data.
Audit Metadata