🤖 refactor: make file edit tool warnings more prominent #502
Conversation
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
Move critical warning to first line with⚠️ emoji for all file edit tools: - file_edit_replace_string - file_edit_replace_lines - file_edit_insert This addresses the issue where agents stream optimistic text about edits succeeding before tool results return. The prominent warning at the start of the description should help agents check results before proceeding with dependent operations like commits, pushes, or builds. Generated with `cmux`
ammar-agent
force-pushed
the
fix-edits
branch
from
c433f05 to
c2d7fa6
Compare
Member
|
Seemd to give a 4% improvement to terminal bench scores |
github-merge-queue
bot
removed this pull request from the merge queue due to failed status checks
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
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Outdated suggestions cannot be applied.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.
Problem
Agents assume file edits succeed and proceed with dependent operations (commits, pushes, builds) before tool results return. During streaming, the agent generates optimistic text like "Now I'll push..." before edit results arrive. When asked later, the agent knows about the failure (it's in context), but the damage is done.
Solution
Move the critical warning to the first line of all file edit tool descriptions with a⚠️ emoji:
Applied to:
file_edit_replace_stringfile_edit_replace_linesfile_edit_insertWhy This Helps
Warnings buried mid-description are easy for streaming models to miss. By placing the warning first with a visual indicator, we increase the likelihood that agents will:
Generated with
cmux