Source Intelligence

DisclaimerUnofficial, and not affiliated with Anthropic. Nearly all of this is read straight out of what ships: npm bundles, captured prompts, published docs. Anthropic's own notes go in verbatim, marked as theirs. The rest is my reading, and every entry carries the strings behind it. If one looks wrong, vote it down and say why.

All of v2.0.9 Home All releases olderv2.0.8 v2.0.10newer
Claude Code v2.0.9

Enhanced Post-Tool-Hook Feedback Handling

What: Feedback from post-tool-use hooks is now properly converted to user messages, making it visible to the AI model for improved context awareness.

How it works now:

  • Post-tool-hook feedback creates messages with subtype "post_tool_hook_feedback"
  • These messages are automatically converted to user messages during processing
  • The AI model can now reference and build upon hook outputs in subsequent responses

Details:

  • Previously, hook feedback was treated as generic "informational" system messages
  • New dedicated functions: eb1() creates the feedback messages, fPB() identifies them
  • Conversion happens in message processing pipeline via AD5() function
  • Includes dedicated UI rendering function h2B() for displaying feedback
  • Evidence: fPB() at line 436505, eb1() at line 437160, conversion logic at line 436513, h2B() rendering at line 405844

See this entry in the whole of v2.0.9 →