deep-research

Pass

Audited by Gen Agent Trust Hub on Feb 28, 2026

Risk Level: SAFENO_CODE
Full Analysis
  • [SAFE]: The skill consists entirely of instructional Markdown content designed to guide an agent through a systematic research process. No executable scripts, binaries, or installation commands are present.
  • [NO_CODE]: The skill does not include any code files, shell scripts, or external package dependencies. All logic is contained within the Markdown instructions.
  • [PROMPT_INJECTION]: The instructions focus on improving research quality and technical accuracy. There are no attempts to bypass safety filters, extract system prompts, or override agent behavioral constraints.
  • [DATA_EXFILTRATION]: No network operation patterns or sensitive file access requests were identified. The URLs referenced in the worked example are reputable and well-known documentation sources.
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 28, 2026, 02:06 PM