prompt: retire autofit-sampler-database from draft/ — completed 2026-08-21 - #258
Merged
Merged
Conversation
…08-21 The task was completed on 2026-08-21 (f58d1e7): record at complete/2026/08/autofit-sampler-database.md, PyAutoFit#1508 closed, 0/9 of the release-run scripts reproduce on current main, no code changed in any repo. That commit deliberately left the prompt file in draft/bug/health_fixes/ on the grounds that it "carries the gate result and the cluster README links it". The first half of that is not true of this file. It was the untouched 2026-07-08 original — 44 lines, no gate annotation, still asserting "The cookbook failures reproduce on current main" and still prescribing "Fix Emcee initialization/ proposal handling". The gate result lives entirely in the record. Its four siblings did receive dated gate annotations (79-148 lines each) and legitimately stay in draft/, because their intermittent parkings cannot be cleared by a single green run; that rule was carried over to this file, which had already crossed into complete/. The effect was not just a stale dashboard row: dashboard.md advertised /start_dev on a body directing an agent to repair a defect proven not to exist. - draft/bug/health_fixes/autofit_sampler_database.md: removed - draft/bug/health_fixes/README.md: row struck through and pointed at the record, matching the aggregator_output_contracts precedent two rows below it - dashboard.md/.html: regenerated (backlog 145 -> 144, bug 33 -> 32) Neither lifecycle.py check nor the intake dashboard's drift detector flagged this: the latter only fires on a `Fix:`-anchored PR reference in the body, and this task shipped no PR. Teaching it to flag a draft prompt whose slug has a complete/ record is left as a separate task. Committed with explicit pathspecs. Co-Authored-By: Claude <[email protected]> Claude-Session: https://claude.ai/code/session_01JcEr8GfbVrMeSQ7iN37gTD
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
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.
What
draft/bug/health_fixes/autofit_sampler_database.mdwas still listed on the dashboard as an unstarted backlog prompt, with a/start_devcommand, despite the task having been completed on 2026-08-21.Why it was still there
The task was completed in f58d1e7: record at
complete/2026/08/autofit-sampler-database.md, PyAutoFit#1508 closed, 0/9 of the release-run scripts reproduce on currentmain, no code changed in any repo. That commit deliberately left the prompt file indraft/bug/health_fixes/, on the grounds that it "carries the gate result and the cluster README links it".The first half of that is not true of this file. It was the untouched 2026-07-08 original — 44 lines, no gate annotation, still asserting "The cookbook failures reproduce on current
main" and still prescribing "Fix Emcee initialization/proposal handling". The gate result lives entirely in the record.Its four siblings did receive dated gate annotations (79–148 lines each) and legitimately stay in
draft/, because their intermittent parkings cannot be cleared by a single green run. That rule was carried over to this file, which had already crossed intocomplete/.The effect was not just a stale row:
dashboard.mdwas advertising/start_devon a body directing an agent to repair a defect proven not to exist.Changes
draft/bug/health_fixes/autofit_sampler_database.md— removeddraft/bug/health_fixes/README.md— row struck through and pointed at the record, matching theaggregator_output_contractsprecedent two rows below it in the same tabledashboard.md/dashboard.html— regenerated viapyauto-brain intake --apply dashboard(backlog 145 → 144, bug 33 → 32)Verification
python3 scripts/lifecycle.py check→ OKautofit_sampler_databasehits remain in either dashboard filecomplete/2026/07/emcee-loguniform-nan.mdalso names the old prompt path but is a sealed historical record, so it was left alone; the four sibling prompts mention it by name in prose only, not as linksdraft/againstcomplete/found this to be the only instance of the driftNot included
Neither
lifecycle.py checknor the intake dashboard's drift detector flagged this — the latter only fires on aFix:-anchored PR reference in the body, and this task shipped no PR. Teaching it to flag a draft prompt whose slug has acomplete/record is left as a separate task.Generated by Claude Code