Tell the LLM that tool outputs are limited and may be truncated#2823
Tell the LLM that tool outputs are limited and may be truncated#2823onmete wants to merge 1 commit intoopenshift:mainfrom
Conversation
The agent system prompt had no guidance about tool output size limits. The LLM had no reason to prefer targeted queries over broad ones that return large amounts of data, contributing to frequent truncation. Add a brief instruction to AGENT_SYSTEM_INSTRUCTION so the LLM plans its tool calls accordingly from the start. Made-with: Cursor
|
[APPROVALNOTIFIER] This PR is NOT APPROVED This pull-request has been approved by: The full list of commands accepted by this bot can be found here. DetailsNeeds approval from an approver in each of these files:Approvers can indicate their approval by writing |
|
@onmete: The following test failed, say
Full PR test history. Your PR dashboard. DetailsInstructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes-sigs/prow repository. I understand the commands that are listed here. |
1 participant
Description
AGENT_SYSTEM_INSTRUCTIONhad no mention of tool output size limits. The LLM had no reason to prefer targeted, scoped queries over broad ones that pull large amounts of data — which is a contributing factor to frequent truncation in practice.Add one sentence to the instruction:
This gives the LLM the context it needs to plan more efficient tool calls from the start, before any truncation occurs.
Type of change
Related Tickets & Documents
Checklist before requesting a review
Testing
make test-unit).make verifypasses (black, ruff, pylint 10/10, mypy clean).AGENT_SYSTEM_INSTRUCTIONout, so they are insensitive to its content.Made with Cursor