intelligence-collection-disciplines
Pass
Audited by Gen Agent Trust Hub on Aug 12, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill establishes an indirect prompt injection surface by instructing the agent to aggregate and analyze data from untrusted external platforms such as social media (Reddit, X), review sites (G2, Trustpilot), and community forums. These sources can be manipulated by third parties to include malicious instructions or biased content intended to mislead the agent.
- Ingestion points: External research signals collected from public web sources specified in the OSINT, HUMINT, and SIGINT disciplines.
- Boundary markers: Includes instructions to follow the autonomous-investigation protocol, which utilizes Fact, Inference, and Assumption labels to categorize incoming data.
- Capability inventory: The skill defines a research framework but does not bundle specific execution scripts or automation tools.
- Sanitization: No explicit data sanitization logic is provided within the skill; it relies on the agent's default processing behavior and the methodology's focus on cross-validation (All-Source Fusion).
Audit Metadata