launchdarkly-experiment-setup
Pass
Audited by Gen Agent Trust Hub on Oct 27, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: No security threats detected. The skill is a legitimate tool for interacting with the LaunchDarkly experimentation platform.
- [PROMPT_INJECTION]: The skill contains no malicious instructions to override agent behavior or bypass safety filters. It includes internal operational constraints that ensure the agent acts only with user consent.
- [DATA_EXFILTRATION]: No sensitive local data access or unauthorized network exfiltration was found. The skill's network activity is restricted to the intended LaunchDarkly MCP server for the purpose of managing experiments.
- [REMOTE_CODE_EXECUTION]: There are no patterns of remote code execution or dynamic code generation. The skill relies on a set of predefined MCP tools provided by the vendor.
Audit Metadata