ai-native-development

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

• Third-party content exposure detected (high risk: 0.90). The skill includes agent tool patterns that fetch and ingest external web content (e.g., references/function-calling.md registers a get_weather tool that fetches https://api.weather.com, and references/agentic-workflows.md plus templates/agent-workflow-template.ts define search_web/fetch_url tools), and those tool outputs are fed back into ReAct and function-calling loops as observations that the agent reads and uses to decide next actions—creating clear exposure to untrusted third-party content that could enable indirect prompt injection.