ab-testing
Pass
Audited by Gen Agent Trust Hub on May 19, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill consists entirely of instructional text and documentation for designing and analyzing experiments. No scripts, binaries, or automated shell commands are included in the skill body.
- [PROMPT_INJECTION]: The 'Mandatory Content Standards' section contains instructions to suppress certain AI behaviors, such as 'No warnings, notes, or disclaimers' and 'No introductory or closing filler phrases.' These instructions are used here as a stylistic persona constraint to ensure high-quality, human-sounding output rather than to bypass safety guardrails.
- [EXTERNAL_DOWNLOADS]: The skill mentions several well-known industry platforms (e.g., Optimizely, VWO, LaunchDarkly) and an online statistical calculator (evanmiller.org). These are provided as informational references for the user and do not involve remote code execution or automated downloads.
- [DATA_EXFILTRATION]: No patterns for accessing sensitive local files or sending data to external endpoints were detected. The skill is purely informational.
Audit Metadata