fix(evals): Restrict rubric judging to visible transcript#855
Merged
Conversation
|
The latest updates on your projects. Learn more about Vercel for GitHub.
|
![cursor[bot]](https://avatars.githubusercontent.com/in/1210556?s=60&v=4){kind=small}
dcramer
force-pushed
the
codex/rubric-visible-transcript
branch
from
a958b54 to
b36e191
Compare
dcramer
force-pushed
the
codex/rubric-visible-transcript
branch
from
b36e191 to
342bf9a
Compare
dcramer
force-pushed
the
codex/rubric-visible-transcript
branch
from
342bf9a to
3160979
Compare
dcramer
force-pushed
the
codex/rubric-visible-transcript
branch
from
3160979 to
375a8a8
Compare
There was a problem hiding this comment.
Cursor Bugbot has reviewed your changes and found 1 potential issue.
❌ Bugbot Autofix is OFF. To automatically fix reported issues with cloud agents, enable autofix in the Cursor dashboard.
Reviewed by Cursor Bugbot for commit 375a8a8. Configure here.
Record observed Slack user messages and direct assistant replies as the visible eval transcript, while keeping collected Slack side effects available for deterministic assertions outside the judge prompt. Document the rubric boundary so semantic checks stay in rubrics and concrete side effects stay in assertions. Co-Authored-By: GPT-5 Codex <codex@openai.com>
dcramer
force-pushed
the
codex/rubric-visible-transcript
branch
from
375a8a8 to
81ed341
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
1 participant
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Slack eval rubrics now judge only the observed user/direct-assistant transcript text. The harness records inbound user messages and direct thread replies into the normalized eval session, while Slack API-captured side effects remain available for deterministic assertions but are excluded from the judge prompt.
The eval docs and policy now make the boundary explicit: use rubrics for semantic visible behavior, and use normal assertions for concrete side effects like tool calls, DB rows, reply counts, and artifacts. Existing core eval rubrics that mentioned reply counts, tool calls, reactions, channel posts, canvas artifacts, or Slack metadata were cleaned up to assert those boundaries deterministically instead.