Skip to content

Conversation

@Tarquinen
Copy link
Collaborator

Summary

  • Adds a new automatic strategy that removes potentially large tool inputs from errored tool calls after a configurable number of turns (default: 4), while preserving error messages for context
  • Zero LLM cost strategy that improves token efficiency in long sessions with errors

Introduces a new automatic strategy that removes potentially large tool inputs
from errored tool calls after a configurable number of turns (default: 4).
Error messages are preserved for context while reducing token usage.

- Add PurgeErrors config interface with enabled, turns, and protectedTools
- Implement purgeErrors strategy with turn-based pruning logic
- Integrate into message transform pipeline
- Update documentation with new strategy details
@Tarquinen Tarquinen force-pushed the feat/purge-errors-strategy branch from 34bd730 to a781513 Compare December 25, 2025 22:05
@Tarquinen Tarquinen merged commit 1fc3709 into dev Dec 25, 2025
1 check passed
@Tarquinen Tarquinen deleted the feat/purge-errors-strategy branch December 25, 2025 22:13
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants