featbit-experimentation-skills

Pass

Audited by Gen Agent Trust Hub on Jun 29, 2026

Risk Level: SAFE
Full Analysis
  • [COMMAND_EXECUTION]: The skill interacts with an external FeatBit MCP server to perform experiment management and feature flag operations. All mutations, such as creating flags, updating targeting, and toggling flag states, are explicitly gated behind user approval requirements.
  • [EXTERNAL_DOWNLOADS]: The documentation references official FeatBit resources and standard open-source Python libraries (numpy, scipy) for statistical analysis. These are well-known and trusted sources.
  • [DATA_EXFILTRATION]: Data operations are confined to the canonical FeatBit API and MCP server. The skill avoids accessing sensitive local files and uses standard authentication headers provided by the environment.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 29, 2026, 01:04 PM
Security Audit — agent-trust-hub — featbit-experimentation-skills