chore(metrics): freeze the proc-v1 process digests - #991
Open
modelmirror wants to merge 2 commits into
Open
Conversation
The freeze cutover: FROZEN_PROCESS_DIGESTS takes the digest of every enabled predictor and evaluator as read off `fedcourts process-digest --all` at the promoted state. The shakedown-emptiness test becomes the frozen-state test: the set must resolve exactly against the current registry, so post-freeze drift in a prompt or registry config surfaces as a mismatch here. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
What
The #907 freeze cutover, as two commits so the record can name the paste:
8bf09ed2pastes the six digests fromfedcourts process-digest --allintoFROZEN_PROCESS_DIGESTS(three predictors, three evaluators, all proc-v1), and converts the shakedown-emptiness test into the frozen-state test — the set must resolve exactly against the current registry, so any post-freeze prompt/config drift fails the suite instead of riding silently.docs/milestones.mdas the cutover marker.Merge with a merge commit, not squash — squashing rewrites the sha the record names.
Pre-freeze evidence
4b93895eand802ea80e), including engine-smoke × claude-code/codex/gemini — real-engine predict cells on the staging keys.gh workflow run integration-test.yml --ref staging -f scenario=engine-smoke -f engine=codex -f docket=73129750(and-f engine=gemini). Recommend running both before merging this.Effect
None until promoted:
is_frozenis read by leaderboard/ops-report from the default branch, so the freeze takes effect at the next (small) staging→main promotion — which closes #907 and the OT2026 first-conference milestone alongside #864.Closes #907 (hand-close on the promotion that carries this, per the freeze-takes-effect-on-main rule).
Merge lane
Maintainer's call — mechanically src+tests+docs, but it IS the pre-registration cutover; leaving the merge to the maintainer deliberately.