Filter out empty messages from LLM prompt construction#3967
Merged
Conversation
Contributor
Contributor
Author
|
@claude iterate to fix the linting and test issues. One of the matrix tests is failing because there is an expect for one attached file that should be checking the count is 2. |
jurgenwerk
approved these changes
Feb 11, 2026
Contributor
There was a problem hiding this comment.
Pull request overview
This pull request filters out empty messages from LLM prompt construction to prevent wasteful and potentially problematic API calls. The change modifies the history construction logic to skip assistant messages that have both an empty body and no tool calls, matching the existing filtering behavior for user messages.
Changes:
- Modified
buildPromptForModelto conditionally add assistant messages only when they have content or tool calls - Added comprehensive test coverage for empty message filtering scenarios
Reviewed changes
Copilot reviewed 2 out of 2 changed files in this pull request and generated no comments.
| File | Description |
|---|---|
| packages/runtime-common/ai/prompt.ts | Updated assistant message construction to filter out empty messages while preserving messages with tool calls |
| packages/ai-bot/tests/prompt-construction-test.ts | Added three new tests to verify empty message filtering for assistant and user messages |
💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.
Empty messages (with no content and no tool calls) could be sent to the LLM, which is wasteful and potentially causes issues. This change: - Skips assistant messages with empty body and no tool calls during history construction (matching the existing guard on user messages) - Adds a safety filter (hasMessageContent) on the final messages array to catch any empty messages from any source, while preserving assistant messages with tool calls and tool response messages CS-10116 https://claude.ai/code/session_01JMzG85yR6uQUXVfnUTegXz
- Remove hasMessageContent filter: it was redundant with the inline guards already present for each message type and would silently mask construction bugs rather than surfacing them - Fix test: use correct field name `id` instead of `toolCallId` for encoded command requests (matching all other tests in the file) - Add tool result assertions to verify the assistant+tool message pairing is preserved when assistant body is empty https://claude.ai/code/session_01JMzG85yR6uQUXVfnUTegXz
lukemelia
force-pushed
the
claude/remove-empty-llm-messages-FNkEO
branch
from
01b3b74 to
f3c72fb
Compare
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.
Empty messages (with no content and no tool calls) could be sent to the
LLM, which is wasteful and potentially causes issues. This change:
history construction (matching the existing guard on user messages)
to catch any empty messages from any source, while preserving
assistant messages with tool calls and tool response messages
CS-10116
https://claude.ai/code/session_01JMzG85yR6uQUXVfnUTegXz