Unverified50% confidenceFactExact time
Since 2023, with the proliferation of AI Agent capabilities such as ChatGPT plugins and Claude tool calling, the practical attack surface of prompt injection attacks has expanded significantly.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
10/11/2026
First Seen
Valid until: 1/9/2027
Sources
Related Entities
Related Claims
Unverified2022年ChatGPT发布后，LLM首次具备了跨任务泛化的能力，使通用型AI Agent从理论走向工程可行69% similarUnverifiedIn the future, new protocols similar to agents.txt may emerge, specifically for declaring a website's policies and capabilities regarding AI Agent access.66% similarUnverifiedIn 2024–2025, major laboratories competed to roll out Agent frameworks, from AutoGPT to Anthropic's Computer Use, with AI shifting from passively answering questions to actively executing tasks.65% similarUnverified自2022年ChatGPT发布以来，具备语义理解与推理能力的基础模型成为Agent的核心引擎64% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/995579API
curl https://kongchang.com/api/v1/knowledge/claims/995579MCP
get_claim(id=995579)