search-specialist

Pass

Audited by Gen Agent Trust Hub on Feb 27, 2026

Risk Level: SAFENO_CODEPROMPT_INJECTION
Full Analysis
  • [SAFE]: No malicious patterns, obfuscation, or unauthorized commands were identified. The skill is limited to natural language instructions for search workflows.- [NO_CODE]: The skill does not contain any executable scripts, code blocks, or external software dependencies.- [PROMPT_INJECTION]: The skill is designed to process content from external websites, which serves as a potential surface for indirect prompt injection. However, this is inherent to its primary purpose as a research specialist.
  • Ingestion points: WebFetch Deep Dive extracts full content from search results.
  • Boundary markers: No explicit delimiters or isolation instructions for external content are provided in the instructions.
  • Capability inventory: Synthesizing information across sources and formulating research findings.
  • Sanitization: No sanitization or validation steps for external content are defined beyond qualitative ranking.
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 27, 2026, 03:35 PM