intelligence-analysis-techniques
Pass
Audited by Gen Agent Trust Hub on Sep 27, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill documentation references external policies and threat modeling guidelines hosted on the vendor's official GitHub repository.
- Evidence: Links to github.com/Hack23/ISMS-PUBLIC found in the compliance mapping section.
- [INDIRECT_PROMPT_INJECTION]: The skill framework involves the ingestion and processing of external OSINT data and parliamentary records. This introduces a surface for indirect prompt injection where malicious content within the analyzed data could attempt to influence the agent's behavior.
- Ingestion points: OSINT collection and SQL database queries described in SKILL.md.
- Boundary markers: No delimiters or specific prompts to ignore embedded instructions are defined in the methodology.
- Capability inventory: Data analysis is performed using Java-based logic and SQL queries, focusing on analytical reporting rather than system-level execution.
- Sanitization: No explicit data sanitization or validation steps are documented for external content.
Audit Metadata