customer-research

Pass

Audited by Gen Agent Trust Hub on Jun 19, 2026

Risk Level: SAFENO_CODE
Full Analysis
  • [SAFE]: The skill consists entirely of instructional markdown and reference templates. No malicious code, hidden commands, or obfuscated content were detected.
  • [DATA_EXFILTRATION]: While the skill references a local context file (.agents/marketing-context.md), it does not perform any network requests, access sensitive system directories, or attempt to harvest credentials.
  • [PROMPT_INJECTION]: The instructions establish a professional persona and enforce high-quality research standards (Jobs-to-be-Done, Ehrenberg-Bass). There are no attempts to bypass safety filters, extract system prompts, or override agent constraints.
  • [REMOTE_CODE_EXECUTION]: The skill does not contain any scripts, package dependencies, or commands to download and execute remote content.
  • [NO_CODE]: The skill is strictly instructional and does not provide any software, scripts, or binary files.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 19, 2026, 12:03 AM
Security Audit — agent-trust-hub — customer-research