Unverified50% confidenceFactExact time
When an agent is able to encode internal state into a URL, the act of reading a webpage is sufficient to transmit sensitive information externally.
1
Sources
50%
Confidence
Long-term
Relevance
9/20/2026
First Seen
Sources
Related Entities
Related Claims
VerifiedIndirect prompt injection refers to attack instructions hidden in external web pages, documents, or API responses accessed by an agent, where the malicious instructions are activated when the agent retrieves and injects them into the context.79% similarUnverifiedagent-readable websites理念希望网站能以结构化、语义清晰的方式向Agent暴露其功能与数据77% similarUnverifiedWebMCP's declarative API includes an agent-invoked boolean attribute that can distinguish whether a form is being filled by an AI Agent or a human, allowing websites to decide whether to skip human interaction steps such as CAPTCHA verification.71% similarUnverified在持有真实登录态的场景下，Agent操作等同于以用户身份在网站上行动，网站无法区分手动操作还是Agent自动化行为68% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/937211API
curl https://kongchang.com/api/v1/knowledge/claims/937211MCP
get_claim(id=937211)