AI article
The README looked harmless. The model saw a hidden instruction.
Community description: How indirect prompt injection reaches an agent through a tool result, and how an open source proxy stops it before the model reads it.
Dev.to | Oct 12, 2026 | yatuk
Automated excerpt
When an agent calls something like read_file("README. md"), the raw text goes back into the conversation as a tool result, comment included. In an agent loop that's the least interesting part of the request. What a tool returns does, on the next request, as a tool result.
Selected automatically from source text; not independently written or fact-checked. Read the original for full context.