creating-agents

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

• Third-party content exposure detected (high risk: 0.80). The skill explicitly includes a Researcher archetype and template (templates/researcher-agent.md) and related references (references/agent-archetypes.md, references/tool-selection.md) that require and instruct agents to use WebSearch/WebFetch to retrieve and synthesize content from public web pages, i.e., untrusted third-party sources that the agent will read and act on.