manage-teams

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

• Third-party content exposure detected (high risk: 0.80). The SKILL.md explicitly instructs the agent to discover and read public, user-created teams via the CLI (e.g., npx @openant-ai/cli teams list --discover and teams get), so untrusted team names/descriptions could be ingested and influence actions like joining or accepting tasks.

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

• Potentially malicious external URL detected (high risk: 0.80). The skill repeatedly runs "npx @openant-ai/cli@latest" at runtime, which fetches and executes remote package code from the npm registry (npx @openant-ai/cli@latest) and is a required dependency for the skill to operate, so it can execute remote code.