deep-research
Pass
Audited by Gen Agent Trust Hub on Jun 23, 2026
Risk Level: SAFEPROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [SAFE]: The skill is a well-designed research utility with built-in safety constraints and iterative verification logic.
- [EXTERNAL_DOWNLOADS]: The skill is distributed via the author's repository on the skills.sh platform, which is a standard distribution method for this ecosystem.
- [PROMPT_INJECTION]: The skill processes untrusted web data as part of its research workflow. This represents a vulnerability surface for indirect prompt injection.
- Ingestion points: Web content retrieved by parallel sub-agents using the Agent tool (SKILL.md Phase 2).
- Boundary markers: No specific delimiters or instructions to ignore embedded commands are defined in the sub-agent prompt contract.
- Capability inventory: The controller manages sub-agents and synthesizes reports, utilizing the Agent tool for external data access.
- Sanitization: The skill relies on a Source-tier ladder and triangulation logic to filter and verify information, rather than traditional string-level sanitization.
Audit Metadata