onboardingV2

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (high risk: 0.70). The skill explicitly instructs the agent to call LaunchDarkly MCP tools and APIs (e.g., "call get-project", "list-feature-flags", "get-flag" in onboarding SKILL.md and mcp-configure/SKILL.md), so it fetches user/third-party project and flag data (user-generated content) and uses those responses to drive decisions and next actions, which creates an indirect prompt-injection risk.

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (high risk: 0.90). The skill instructs the agent to fetch and install remote skill packages at runtime (e.g., via npx/plugin install from the LaunchDarkly repo https://github.com/launchdarkly/ai-tooling), which will download and execute code that directly controls agent prompts/workflow, and those installs are required for the onboarding flow—this meets the criteria for a runtime-executed external dependency.