Extreme agent looping: 30-50 tool calls consuming 400K-700K+ tokens across multiple models#558
Draft
langsmith-engine-dev[bot] wants to merge 1 commit into
Draft
Conversation
…ooping Traces show the agent making 30-50 tool calls with repeated identical queries, consuming 400K-700K+ tokens. The recursion_limit of 100 allows far too many iterations. A limit of 25 (~10-12 tool call rounds) is sufficient for any legitimate question while preventing runaway loops.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Multiple traces show the agent making 30-50 tool calls with repeated identical queries, consuming 400K-700K+ tokens and taking 150-300+ seconds. This is a more severe variant of the Gemini looping pattern and affects multiple models including GPT-5 Mini. The worst case (019d73be) made 50 tool calls repeatedly searching "composite", "evaluator", and "LangSmith Evaluation" for a Korean-language question about annotation queues. The recursion_limit in langgraph.json is set to 100, which is far too high to prevent this. Reducing it to 20-30 and adding tool call deduplication would prevent these extreme cases.