onboardingV2

Pass

Audited by Gen Agent Trust Hub on May 2, 2026

Risk Level: SAFE
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill downloads and installs companion agent skills (e.g., mcp-configure, sdk-install, launchdarkly-flag-create) from the vendor's official GitHub repositories (launchdarkly/experimental-agent-skills and launchdarkly/agent-skills).
  • [COMMAND_EXECUTION]: Executes shell commands to detect project structure, check port availability (e.g., lsof), and install SDK dependencies using native package managers (npm, pip, dotnet, etc.).
  • [CREDENTIALS_UNSAFE]: Safely manages LaunchDarkly SDK keys and API access tokens. It uses mandatory decision points (D7 and D4-LOCAL) to ensure the user explicitly approves how secrets are stored, recommending .env files and verifying they are listed in .gitignore to prevent credential leakage.
  • [SAFE]: The skill implements experience detection to adjust its guidance level and follows strict sequence enforcement to ensure security configurations, such as MCP authorization, are verified before proceeding.
Audit Metadata
Risk Level
SAFE
Analyzed
May 2, 2026, 03:45 PM