Publish the work the bees already pushed - #4322
Merged
Merged
Conversation
401 queen-* branches on the remote and the last bee pull request opened on 2026-09-17. Seventeen pull requests merged in the last day, none from a bee. Of the forty most recent branches 24 carry real commits, 9 carry only a salvage commit, 7 are empty - and all 24 were invisible, because nothing in this system opens a pull request. The Queen skips issues that ARE pull requests and never creates one, the container holds no push credential by design, and the job that used to do it was a person running a script on a laptop. Twenty bees at ninety percent utilisation for three days, shipping nothing, while every instrument reported health - because every instrument measured dispatch and none measured delivery. publish.py opens them, five per run, twice an hour. Five because each pull request starts twenty-odd workflow runs and this repository has already had a queue of 1073 superseded runs starve its own gates. It adds the one thing a bee cannot: a docs/now entry, with `Closes #N` in the COMMIT message, because L1 TRACEABILITY reads commit messages and not pull request bodies. It refuses a branch that already has a pull request, that has no commits or no diff, whose issue is closed, that wrote outside its `## Boundary`, that conflicts with master (GitHub runs no required check on a branch it cannot merge, so it reports as waiting for CI forever), or that is too large to be one turn - queen-3678 sits 3269 commits ahead. Proven before shipping: the probe #4317 needed exactly that entry and failed L1 until the commit said Closes. Dry run: 6 would publish, 1 conflicts, 1 too large, 8 issues closed, 2 no commits, 1 already had a pull request. CENSUS, moved by the new workflow: quiet 58 -> 59 workflow files and 137 -> 139 "named a path but not quiet"; shell 58 -> 59 files, 79 -> 80 jobs, 269 -> 272 run-steps, 248 -> 251 run with `bash -eo pipefail`. Classified in check_pr_branch_filters.py in this same commit; the ceiling stays at 24. Closes #4321 Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
gHashTag
enabled auto-merge (squash)
September 20, 2026 12:01
Contributor
|
📓 NotebookLM Notebook linked to this PR
This notebook contains session context, decisions, and artifacts for this work. |
Contributor
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.
Closes #4321
Twenty bees, three days, nothing shipped
Of the forty most recent branches: 24 carry real commits, 9 only a salvage
commit, 7 empty. All 24 invisible.
Nothing in this system opens a pull request. The Queen skips issues that
are pull requests and never creates one; the container holds no push
credential by design; the job that used to do it was a person running a script
on a laptop. Every instrument reported a healthy swarm at 90% utilisation —
because every instrument measured dispatch, and none measured delivery.
What it refuses, and why each
## Boundaryqueen-3678sits 3269 commits aheadIt adds the one thing a bee cannot
A
docs/now/entry, withCloses #Nin the commit message — L1TRACEABILITY reads commit messages, not pull request bodies. The entry says the
publisher wrote it rather than pretending a bee did.
Measured before shipping
A probe published by hand — #4317,
queen-4286, a restoredswap()and a test— needed exactly that entry and failed L1 until the commit said
Closes. Then:Both ledgers a new workflow moves are updated in this same commit: classified in
check_pr_branch_filters.py(ceiling stays 24) and the census re-blessed withevery moved number named.
🤖 Generated with Claude Code