codex-bot-andresmgsl/ceremony
1
0
Fork 0
forked from heavy-duty/ceremony
Code Pull requests Activity
ceremony/docs/CONSUMERS.md
Andriujose d87d76d64b fix: align marker guard with release oracle
2026-08-03 22:56:02 +00:00

783 lines
41 KiB
Markdown
Raw Blame History

# Consumer setup
How a repo adopts the ceremony — bootstrap for a greenfield repo, a
conversion checklist for a repo carrying its own copy of the machinery,
and the policies that keep either honest afterwards. The doctrine (what a
release *is*, the doors, the guards, the drill) lives in
[../README.md](../README.md); this guide is the how-to. It is meant to be
sufficient on its own: a conversion executed from this guide should need
zero out-of-band knowledge, and gaps found while converting are filed as
edits to this guide (#12).
## Prerequisites
- **Repo shape**: work lands on a `main` default branch by PR; fork PRs
are fine — the merge door rides `push` to `main`, never `pull_request`
([release.yml](../.github/workflows/release.yml#L70-L74), box#97), and
the label read goes through the API
([lib/facts.sh](../lib/facts.sh#L88-L101)), so the ceremony never needs
the PR's own context. No PAT, no secrets: every permission the flow uses
is the caller-declared `GITHUB_TOKEN` grant.
- **Pick the version backend**: `file` (a `VERSION` file — box, rig,
incubator) or `package-json` (the `version` field, lockfile kept in sync
on the post-release bump — cast). This is the workflow's one input; the
full configuration surface of the ceremony is enumerated in
[#1](https://github.com/heavy-duty/ceremony/issues/1) ("The
configuration axes").
- **The `release` label must exist** before the first ceremony PR — it is
the merge door's declared-intent read
([lib/facts.sh](../lib/facts.sh#L88-L101)). Bootstrap it via the labels
sweep caller's `workflow_dispatch`
([Labels automation](#labels-automation)), or create it by hand,
matching the core table
([actions/labels-reconcile/labels-reconcile.sh](../actions/labels-reconcile/labels-reconcile.sh#L369)):
```sh
gh label create release --color 0E8A16 \
--description "Release flow and version/packaging work"
```
## Bootstrap a new repo
The greenfield path (incubator's, #16) — the repo never owns a copy of
the machinery at all:
1. **`VERSION` at `X.Y.Z-dev` — never bare.** A first version that never
carried `-dev` hits the decide table's refuse row and has to ship by
the tag door (the known first-release edge, cast#111;
[lib/decide.sh](../lib/decide.sh#L70-L74)). Bootstrapping at `-dev`
keeps the repo clear of it entirely. (`package-json` backend: the
`version` field, same rule.)
2. **An armed changelog: a `CHANGELOG.md` preamble plus `changelog.d/`.**
The changelog file starts as preamble only — no section; the first
release writes the first one. The fragments directory beside it is the
arming (#112): it carries a `README.md` marker naming the assembler and
the doctrine — take ceremony's own
[changelog.d/README.md](../changelog.d/README.md) at the pin — which is
what keeps the directory tracked while it holds no fragments and what
`changelog-armed` asserts. Every behavior-change PR then writes
`changelog.d/<issue>.md` ([The changelog rule](#the-changelog-rule));
the release PR assembles the section
([Assembling a release section](#assembling-a-release-section)).
Fragment mode is available at `0.2.0` and later, and not in `0.1.0`.
A consumer pinned to `0.1.0` bootstraps the legacy shape instead — the
preamble plus an empty `## Unreleased` section for entries to land
under — and converts on the pin bump to `0.2.0` or later; never mix
refs to adopt it early.
3. **`drills/README.md`** defining what a drill *means* in this repo —
each repo names its own
([the drill doctrine](../README.md#the-drill-doctrine)). Plain
`drills`, not a dot-directory
([drill-recorded.sh](../actions/drill-recorded/drill-recorded.sh#L49-L52)).
4. **`.github/workflows/release.yml`** — the caller, verbatim from
[Release workflow](#release-workflow) below.
5. **CI guard steps** in the repo's `ci.yml`:
```yaml
- uses: actions/checkout@v4
with:
# changelog-monotonic and changelog-assembled compare HEAD
# against the merge base; a checkout that cannot resolve it is
# a hard failure in CI, not a skip (a guard that can quietly
# stop guarding is the failure shape these checks exist to
# refuse).
fetch-depth: 0
- uses: heavy-duty/ceremony/actions/changelog-armed@<pinned-tag>
- uses: heavy-duty/ceremony/actions/changelog-monotonic@<pinned-tag>
# changelog-assembled is available at 0.2.0 and later, not in
# 0.1.0. Adopt this step with the pin bump to 0.2.0 or later;
# never mix refs. Green NOTICE on every non-release PR; on a
# release PR it asserts the stamped section is exactly the
# fragments it consumed.
- uses: heavy-duty/ceremony/actions/changelog-assembled@<pinned-tag>
- uses: heavy-duty/ceremony/actions/drill-recorded@<pinned-tag>
# runner-isolated is available at 0.2.0 and later, not in 0.1.0.
# Adopt this step with the pin bump to 0.2.0 or later; never mix
# refs.
- uses: heavy-duty/ceremony/actions/runner-isolated@<pinned-tag>
```
`changelog-armed` and `drill-recorded` take
`version-source: package-json` where that is the backend; every guard's
inputs and defaults are in its `action.yml`
([actions/](../actions/)). Adopting the agent team flow adds the
`docs-sync` step ([below](#adopting-the-agent-team-flow)).
`runner-isolated` asserts that no `pull_request`-triggered workflow
names a self-hosted runner — a PR workflow runs the branch's code, and
unreviewed fork code must never execute on your own hardware
([#58](https://github.com/heavy-duty/ceremony/issues/58)). It fires on
the PR that first mixes a PR trigger and a self-hosted `runs-on` in
one file; the unblock is splitting the workflow. A repo with **no**
self-hosted runner still wants it: the guard's value is the day
somebody adds one.
This guide documents `main`. A marker is the literal token
`**unreleased**` immediately followed by its issue citation (for example,
`(#238)`); whitespace between them may include a line break. A citation is
mandatory, because a marker the guard cannot trace is a marker it cannot
prove false. A marker for this repository's own issue uses bare `#N`.
Cross-repo citations such as `(crew#293)` satisfy the traceability rule but
are not compared with this repository's release section. The ceremony-only
`marker-check.sh` guard enforces these rules. The release PR that ships the machinery clears, in that same PR,
every marker its own assembled section makes false: the section cites its
issues, each marker cites the same issue, and the release PR's diff is the
one place both halves are visible at once (#221). If an action does not exist at the
consumer's pinned tag, adopt it with the pin bump to the first tag that
carries it; never mix a moving or newer ref into an otherwise exact-pin
consumer. In particular, `0.1.0` carries `changelog-armed`,
`changelog-monotonic` and `drill-recorded` plus `docs-sync`, but not
`changelog-assembled` or `runner-isolated`.
6. **`.github/workflows/refs-guard.yml`** — the body-aware guard is its own
caller because `edited` is load-bearing: #200 gained its accidental
closing keyword after the PR opened, with no push to wake ordinary CI.
It costs the consumer one read-only workflow file and no other machinery:
```yaml
name: Refs guard
on:
pull_request:
types: [opened, edited, reopened, synchronize]
permissions:
contents: read
pull-requests: read
jobs:
refs-not-closing:
runs-on: ubuntu-latest
steps:
- uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683 # v4.2.2
- uses: heavy-duty/ceremony/actions/refs-not-closing@<pinned-tag>
```
`refs-not-closing` is **unreleased** (#218) until the first tag carrying it.
Adopt this caller with that ordinary pin bump; never point only this file
at a moving or newer ref.
7. **Labels automation** (optional but recommended): the two callers from
[Labels automation](#labels-automation) — the event-facing labels
caller and the sweep caller (#209) — plus `.github/labels.conf`
(panel + the repo's `scope:*` rows) and `.github/labeler.yml` (the
path→scope globs). Run the sweep caller's `workflow_dispatch` once —
**this bootstraps the taxonomy, `release` label included** — and use it
again whenever an operator needs a full-board sweep immediately.
8. **The artifact hook** (optional): `.github/actions/release-artifact/`
per [The artifact hook](#the-artifact-hook). No hook → the source
tarball is the package.
From there the flow is the doctrine: ordinary PRs write their fragment,
the ceremony PR makes
[the three stamps](../README.md#what-a-release-is), a human merges, the
machine transcribes.
## Convert an existing repo
The box/rig/cast path — the repo carries its own copy of the machinery
and hands it over. The conversion PR is release-flow work: label it
`release` if the repo's conventions ask for that, and either way it lands
as a green `NOTICE` no-op on main — the decide table's green rows exist
precisely so the machinery is safe to work on
([lib/decide.sh](../lib/decide.sh#L6-L12)).
- [ ] Replace `.github/workflows/release.yml` with the caller from
[Release workflow](#release-workflow) — **whole file**, keeping its
load-bearing comments. Check the result has **one** `push:` key
carrying both filters: YAML maps are last-key-wins, and a second
sibling `push:` silently kills a door (rig's review catch).
- [ ] Swap the guard *script* steps in `ci.yml` for the `uses:` steps in
the bootstrap list above (with `fetch-depth: 0` on the checkout).
- [ ] Add `refs-guard.yml` from the bootstrap list with the same ceremony
pin as the release caller and CI guard steps.
- [ ] Replace `labels.yml` with the caller from
[Labels automation](#labels-automation) and add the sweep caller
`labels-sweep.yml` beside it (#209); extract
`.github/labels.conf` from the old reconciler's embedded config —
the `panel=` roster line and the repo's `scope:*` rows
([the format](#labels-automation)). `.github/labeler.yml` stays as
it is (path globs are inherently repo-specific).
- [ ] Convert the changelog to fragments (requires a pin at the first tag
carrying fragment mode — not `0.1.0`): move every entry under
`## Unreleased` to `changelog.d/<issue>.md`, verbatim — the filename
is derivable from the entry's own `(#N)`; an entry citing several
issues goes to the file for the first cited — delete the
`## Unreleased` heading, and add the `changelog.d/README.md` marker
([bootstrap step 2](#bootstrap-a-new-repo)). Published sections stay
byte-identical; `changelog-monotonic` proves that on the conversion
PR, and `changelog-armed` refuses a surviving `## Unreleased` the
moment the directory exists. Rewrite the repo's own contributor
docs that say "add a line under `## Unreleased`" in the same PR —
split either way, main lies for as long as the split lasts.
- [ ] Delete the now-shadowed copies — zero shared scripts remain:
`.github/scripts/release-notes.sh` (box, cast) or
`release-lib.sh` (rig), `changelog-armed.sh` (box),
`changelog-monotonic.sh`, `drill-recorded.sh`,
`labels-reconcile.sh`.
- [ ] Trim the repo's test suite to repo-specific tests: the machinery
tests go — they live in this repo's `test/` now, run by its CI —
while the repo's own surfaces stay (box/rig's install-channel halves
of `test/release.sh`, cast's `install-sh` tests). A machinery test
*file* goes whole when its subject moved (rig's
`test/labels-reconcile.sh` sourced the deleted reconciler), and so
do tests that pin the old workflow's shape — a grep or awk against
`release.yml`/`ci.yml` internals fails against the caller stub, not
because the stub is wrong (rig #13's conversion).
- [ ] Sweep the repo's other docs for pointers at the deleted paths —
`drills/README.md` and any labels doc typically cite the old
`.github/scripts/*.sh` by path; repoint them at the pinned actions.
A repo carrying its own copy of a doc the mirror vendors (rig's
root `LABELS.md`) retires it in the same PR: a hand-maintained
copy beside a machine-verified mirror is the drift the mirror
exists to end.
- [ ] Shrink CONTRIBUTING's release section to a pointer at
[this repo's README](../README.md) plus what is genuinely per-repo:
the drill meaning (`drills/README.md`), artifact notes, the
changelog house style if it differs from
[the portable rule](#the-changelog-rule).
- [ ] What stays, per repo, forever: `VERSION` (or the `package.json`
version), `CHANGELOG.md`, `changelog.d/`, `drills/`, `.github/labeler.yml`,
`.github/labels.conf`, the optional
`.github/actions/release-artifact/` — the full kept-vs-moved table
is in [#1](https://github.com/heavy-duty/ceremony/issues/1).
## Release workflow
The reusable release workflow implements both doors of the ceremony — the
merge door (merging the `release`-labeled ceremony PR ships it) and the tag
door (a bare `X.Y.Z` tag push as the manual fallback and backfill). The
design essay lives in the workflow's own header comment; the doctrine in
issue #1.
The consumer's **entire** `release.yml`:
```yaml
name: release
# Triggers and permissions MUST live here (a called workflow cannot define them):
on:
# ONE push key, both filters — YAML maps are last-key-wins; a second sibling
# `push:` silently replaces the first and kills a door (rig's review catch).
push:
tags: ["**"] # every tag — a wrong tag must FAIL the assert loudly,
# never be skipped by a shape filter that didn't match
branches: [main]
permissions:
contents: write # tag ref create + release create + the bump push
pull-requests: write # decide's label read; the bump-fallback `gh pr create`
issues: write # --label on that fallback PR rides the issues API
jobs:
release:
uses: heavy-duty/ceremony/.github/workflows/release.yml@<pinned-tag>
with:
version-source: file # or: package-json
```
`version-source` is the only input: `file` (a `VERSION` file — box, rig,
incubator) or `package-json` (the version field, lockfile kept in sync on
the post-release bump — cast). Everything else a repo might vary is a change
to the ceremony itself, made in this repo, once.
Keep the merge door on `push` to `main` — never `pull_request`: a
`pull_request` run from a public fork gets a read-only `GITHUB_TOKEN` that
`permissions:` cannot raise (box#97), and every ceremony PR in this org is
cross-repo from a bot fork.
Bootstrap the version at `X.Y.Z-dev`, not bare: a first version that never
carried `-dev` hits the decide table's refuse row and has to ship by the
tag door instead (the known first-release edge, cast#111).
### The artifact hook
If the repository contains `.github/actions/release-artifact/action.yml`,
both doors invoke it — after the tag exists, before `gh release create` —
with the release `version` as input and `RELEASE_ASSETS_DIR` exported.
Contract for hook authors:
- Drop finished files into `$RELEASE_ASSETS_DIR`; every file there is
uploaded as a release asset.
- Exit non-zero to abort the release.
- The hook owns its own toolchain (checkout is done; install node, docker,
whatever it needs, itself).
A failed hook leaves the tag created but no release published. Recovery is
the tag door's semantics: fix the cause, then delete and re-push the same
tag (the tag door publishes for it), or run `gh release create` by hand from
a fixed tree. The merge door's nothing-exists assert will refuse a re-run of
the completed merge, by design.
No hook → no assets: for a pure-bash tree, GitHub's source tarball for the
tag IS the package. Worked examples land with the conversions: cast's tgz
build (#15) and incubator's GHCR image push (#16).
## Labels automation
The labels automation is two reusable workflows since #209, adopted
together at the same pin:
- **`labels.yml`** — the event-facing half, called on PR and issue events.
Two jobs: additive path-based `scope:*` labels, and a few-seconds
`trigger` job that wakes the sweep by dispatching the consumer's sweep
caller (`gh workflow run`, plain `GITHUB_TOKEN` — `workflow_dispatch` is
one of the two documented exemptions from the token's no-retrigger rule,
so no PAT anywhere in the path and no loop: the sweep dispatches
nothing).
- **`labels-sweep.yml`** — the reconcile sweep: PR state, blockers,
handoff, stale status, the issue work queue, and the `needs-ruling`
invariants on both surfaces — the bare-flag check and the 7-day
comment-only nudge (#52; the sweep reads that flag and never writes it).
Detached from PR-triggered runs on purpose: all sweeps serialize through
one shared concurrency group, and GitHub records every queue-displaced
run as CANCELLED — harmless (the surviving sweep does its work) until it
rode a `pull_request_target` run and the ❌ landed on that PR's checks
as fake red CI that GitHub refuses to rerun (crew#250: `gh run rerun`
and its `--failed`/`--job` forms all decline a queue-displaced run).
Behind its own caller, a displaced sweep cancels on the
Actions tab, attached to no PR; PR checks show `scope` and the green
`trigger` only.
The consumer keeps its path mapping in `.github/labeler.yml` and its
review panel plus scope taxonomy in `.github/labels.conf`.
**Additive means additive** (available at `0.3.0` and later — #130): the
scope job's only label
write is `POST /issues/{n}/labels`, which adds the derived scopes and removes
nothing, so a label applied while the job runs survives it. Earlier tags used
`actions/labeler@v5`, which — even under `sync-labels: false` — replaces the
whole label set and silently drops a label written mid-job (ceremony#128 lost
its `release` that way). With the same pin bump, `.github/labeler.yml` keeps
its format but the accepted shape becomes exactly the one this guide has
always shown: label → `changed-files` → `any-glob-to-any-file`, block or flow
style, globs over `**`, `*` and `?` (`**` crosses `/`, the others do not; the
whole path must match). Any other labeler key — `all-globs-to-all-files`,
branch matchers, negations — fails the run loudly instead of being
half-honoured. The reconcile sweep also warns (never sets) when a non-draft
PR carries a bare `X.Y.Z` version differing from its base but no `release`
label — the merge door would refuse that merge, and the sweep says so first.
The complete event-facing caller is:
```yaml
name: labels
on:
pull_request_target:
# Fork PRs; these carry the head/draft/review facts state:* derives from.
# labeled/unlabeled are the handoff wake (state:needs-human confirmed here);
# synchronize re-derives on every push. review_requested/review_request_removed
# (shipped in 0.3.0, ceremony#137) wake the sweep that clears
# blocker:unrequested when the panel is asked.
types: [opened, reopened, ready_for_review, converted_to_draft, synchronize, labeled, unlabeled, review_requested, review_request_removed]
# Available at 0.2.0 and later (the first tag carrying ceremony#32); a
# consumer pinned to 0.1.0 omits this block.
issues:
# Narrowed (#199) to the actions carrying a queue-state change the hourly
# cron cannot wait one cadence for: opened → the mint→needs-triage check,
# closed → the blocker-closes→ready self-heal, edited → a body rewrite of the
# `Blocked by #N` declaration the sweep parses, reopened → a closed issue
# re-entering the queue. Dropped: labeled/unlabeled/assigned/unassigned —
# validation + the 48h claim clock, caught within one cadence, and
# labeled/unlabeled were the issues-churn source. The handoff wake is
# pull_request_target:labeled, not issues, so this leaves it intact.
types: [opened, closed, edited, reopened]
permissions:
contents: read
checks: read # mergeability/check-rollup read for PR state
statuses: read # commit-status rollup read for PR state
actions: write # the trigger job's `gh workflow run` dispatch of the sweep caller (#209)
issues: write
pull-requests: write
jobs:
labels:
uses: heavy-duty/ceremony/.github/workflows/labels.yml@<pinned-tag>
# If the sweep caller below is named anything but labels-sweep.yml,
# say so: `with: { sweep_workflow: <filename> }`. Ceremony's own
# dogfood does (self-labels-sweep.yml).
```
And the complete sweep caller, `labels-sweep.yml` beside it — the hourly
cron lives HERE since #209, not on the labels caller:
```yaml
name: labels-sweep
on:
# The consumer owns this cadence (#203). Hourly is the recommended default
# when no other engine drives board state: the cron is then the sweep's only
# wake for four transition classes — a review verdict landing (no
# pull_request_review trigger on the labels caller), blocker:ci-red
# set/cleared, blocker:conflict when another PR merges under this one, and
# time-based stale / 48h claim-reclaim. The labels caller's events carry the
# rest in seconds, one trigger-job dispatch away. Hourly trades ≤1h of
# latency on those four while cutting nominal scheduled sweeps from four an
# hour to one at GitHub's 1-minute floor. Do not delete the cron: it is their
# discovery path. If another engine writes some of those transitions, only
# the classes with no other writer bound the cadence; relax it only as that
# list shrinks.
schedule: [{cron: "0 * * * *"}]
# A manual full-board sweep. A bare dispatch (input default "yes") also
# bootstraps the taxonomy on a fresh repo. The labels caller's trigger job
# wakes this workflow with bootstrap=no on every board event, so the
# declared input is part of the contract: a dispatch naming an undeclared
# input is refused, and the trigger job goes loudly red.
workflow_dispatch:
inputs:
bootstrap:
description: Bootstrap the label taxonomy before sweeping
type: choice
options: ["yes", "no"]
default: "yes"
permissions:
contents: read
checks: read # mergeability/check-rollup read for PR state
statuses: read # commit-status rollup read for PR state
actions: read # workflow-run nodes inside the check rollup — private repos do not imply it (incubator#60)
issues: write
pull-requests: write
jobs:
sweep:
uses: heavy-duty/ceremony/.github/workflows/labels-sweep.yml@<pinned-tag>
# If this repo's PR-facing labels caller is named anything but `labels`,
# pass that name: `with: { pr_workflow_name: <name> }`. The sweep exports
# it as SELF_WORKFLOW so the label machinery's own check entries (scope,
# trigger) never count toward blocker:ci-red — a red trigger means "fix
# the caller", which no PR edit can do (#208 reads it).
```
Naming any permission sets every unnamed permission to `none`. Public
repositories allow check data to be read regardless, but a private consumer
needs the explicit reads above; without them the failure appears as an empty
`state:*` axis on the board rather than a red workflow run. The labels
caller's `actions: write` is different — it is required everywhere, public
repos included: the trigger job's `gh workflow run` is a write, and without
it every event run goes red at the trigger.
**The failure mode to know before bumping**: a consumer that bumps its pin
to a #209-carrying tag without adding the sweep caller keeps green-looking
silence nowhere — the trigger job goes **red on every PR and issue event**
(workflow-not-found; likewise on a sweep caller missing its `bootstrap`
input, or a labels caller missing `actions: write`), and event-woken sweeps
stop until the caller lands. That loudness is deliberate: never read
silence, or a green `scope` alone, as health. Make the adoption one atomic
PR — pin bump, sweep caller file, `actions: write` line together.
The `issues:` trigger is available at `0.2.0` and later — `0.2.0` is the
first tag carrying ceremony#32. A consumer pinned to `0.1.0` omits it. Adopt
it only by bumping every ceremony reference to `0.2.0` or later; never mix
refs to adopt it early. The type list has grown then narrowed across tags:
`0.2.0` (ceremony#32) shipped `[opened, labeled, unlabeled, assigned,
unassigned, closed]`; `0.3.0` (ceremony#144) added `edited` and `reopened`;
ceremony#199 narrows it to `[opened, closed, edited, reopened]` and relaxes the
cron to hourly, so a consumer picks up the smaller trigger surface at the pin
bump to the first tag carrying ceremony#199. The narrowing drops
`labeled`/`unlabeled`/`assigned`/`unassigned` — validation and the 48h claim
clock, which the hourly cron catches within one cadence, and `labeled`/
`unlabeled` were the issues-churn source — while **keeping** #144's `edited`/
`reopened`: those carry a queue-state change an event uniquely carries (a body
rewrite of `Blocked by #N`, and a closed issue re-entering the queue), so the
must-fail in ceremony#199 keeps them on events. `opened` drives the
mint→`needs-triage` check and `closed` the blocker-closes→`ready` self-heal;
the stub and ceremony's own caller stay byte-for-byte identical, the parity
#144 established.
The two-caller split (ceremony#209) is available at `0.4.1` and later. A
consumer pinned to `0.4.0` or earlier keeps the previous single-caller
shape — the labels caller carrying the cron, `workflow_dispatch`, and
`actions: read` — and adopts the split at the pin bump to `0.4.1` or
later. Never mix refs to adopt it early.
The migration is **one atomic PR** with exactly four edits — crew, the
consumer whose displaced-check evidence drove #209 (crew#227, crew#250),
is the worked example; written here against `0.4.1`, the first tag
carrying the split:
1. **Pin bump, every reference together** ([Version pinning](#version-pinning)):
`0.4.0` → `0.4.1` in the labels caller's `uses:` line **and in every
other ceremony `uses:` in the repo** — crew also pins in
`release.yml` and its `ci.yml` guard steps. A repo on the doctrine
mirror re-runs `docs-sync --fix` in the same PR.
2. **New file `.github/workflows/labels-sweep.yml`** — the sweep caller
stub above, verbatim, `bootstrap` input included (the trigger's
`-f bootstrap=no` dispatch is refused if the input is undeclared).
3. **The hourly cron RELOCATES — it is moved, never copied.** Delete the
`schedule:` block (and the bare `workflow_dispatch:`) from the labels
caller in the same edit that adds the sweep caller.
**Warning**: a consumer that copies the sweep caller and leaves the
old schedule on the labels caller gets DOUBLE sweeps — every cron tick
fires both callers into the one shared `labels-reconcile` group — so
displacement goes **up**, and the fix reads as the bug getting worse.
4. **`actions: write` on the labels caller** — consumers carry
`actions: read` today (crew does); the trigger job's `gh workflow run`
is a write. The sweep caller keeps `actions: read`.
Bump without the sweep caller and the trigger job goes red on every PR
and issue event — the loud failure mode above — so never split these
four edits across PRs.
`pull_request_target` is intentional: fork PRs need the base repository's
token to write labels. The reusable workflows execute no PR code. They check
out only the consumer's base branch and the pinned ceremony implementation.
The #52 ruling invariants ride exactly these triggers — but the caller above
is no longer the #18 shape, so adopting current triggers is a stub edit, not
a bare pin bump. `review_requested` and `review_request_removed` on
`pull_request_target:` shipped in `0.3.0` (ceremony#137) — the wake that
clears `blocker:unrequested` the moment the panel is asked, without which a
quiet repo wears that flag until the backstop cron; a consumer picks them up
by pinning `0.3.0` or later, never through mixed refs.
`.github/labels.conf` has one mandatory panel setting, one mandatory
`triage-actors` setting, zero or more optional per-author panel rows, and
then zero or more scope rows:
```text
panel=claude-bot example-codex-bot example-grok-bot
panel[example-builder]=example-codex-bot example-grok-bot
triage-actors=example-triage-bot
scope:cli|C5DEF5|The command-line surface
scope:docs|C5DEF5|Documentation
```
The mandatory `triage-actors=` setting is likewise accepted at `0.2.0` and
later, and not by `0.1.0`. At that tag the file contains `panel=` plus scope
rows only; adding `triage-actors=` is a parse failure, not an ignored setting.
Add it at the same pin bump as the `issues:` trigger — `0.2.0` or later —
never before it and never through mixed refs.
The optional `panel[<login>]=` rows are available at `0.5.0` and later (#224). A row names
the effective panel for PRs authored by exactly that login — the reconciler
computes that PR's required set from the row, minus the author as always —
and every other author keeps the base `panel=`, which stays mandatory. The
panel is configured or it is the base one: ceremony never infers a reviewer
set from the model behind a login. On any earlier pin a bracketed row is a
**parse failure, not an ignored setting** — the same shape `triage-actors=`
bought at `0.2.0`, but harsher in practice: the reconcile job dies on every
PR event and every sweep until the row is removed, so the whole label board
goes down. Add the row only at or after the pin bump that carries it, never
before it and never through mixed refs.
Both actor lists are whitespace-separated. `triage-actors` names the identities
allowed to mint issues without the sweep applying `needs-triage`. Label rows use exactly
`name|color|description`; blank lines are ignored and extra pipes are refused.
There are no comment lines: every non-blank line must be the `panel=`
setting, a `panel[<login>]=` row, the `triage-actors=` setting, or a label
row, so `#`-prefixed prose is a parse failure, not a comment (rig #13's
conversion found this the hard way — keep the file data only).
Core state, blocker, work-queue, and release labels come from ceremony. Scope
rows remain consumer-owned because paths and surfaces differ by repository.
After adding the callers and configuration, dispatch the sweep caller once
to bootstrap labels on a fresh repository. A bare dispatch is also the
operator's general manual full-board sweep — the answer when the board
looks wrong now rather than after the next scheduled cadence:
```sh
gh workflow run labels-sweep.yml -R <owner>/<repo>
```
Ceremony dogfoods the callers under the filenames `self-labels.yml` and
`self-labels-sweep.yml`, so the equivalent command in this repository
substitutes that filename. Scheduled and trigger-driven runs only
reconcile; they do not repeatedly upsert the taxonomy (the trigger's
dispatch carries `bootstrap=no`). When a ceremony pin bump adds a core
label, bump the pin first and then re-dispatch; the scheduled sweep warns
when the pinned taxonomy declares a core label the repository lacks.
## Doctrine mirror
Machinery is consumed by reference — GitHub fetches the workflows and
actions above from the pin at run time — but documents have no runtime: an
agent reads the working tree it stands in. So the agent-facing doc set
declared by ceremony's `docs/VENDORED.txt` is vendored into each consumer at **`.ceremony/`**,
byte-identical to ceremony at the pin, plus a generated `.ceremony/README.md`
marking the directory machine-managed. `actions/docs-sync` owns the copy:
`--fix` writes it (and deletes what the manifest dropped — mirror means
mirror), `--check` re-diffs it in CI on every PR, so a hand edit or a stale
pin goes red instead of quietly governing.
`RELEASES.md` joins that mirror with the first tag carrying ceremony#248.
It is **unreleased** (#248) until that tag exists: consumers add
`.ceremony/RELEASES.md` only with the ordinary pin bump and re-sync, never by
copying it ahead of their pinned doctrine set.
The consumer's ci.yml gains the guard alongside the others:
```yaml
- uses: actions/checkout@v4
- uses: heavy-duty/ceremony/actions/docs-sync@<pinned-tag>
```
`mode` defaults to `check`. There is no ref input: the action reads the pin
from the consumer's own `.github/workflows/release.yml` — the same single
`uses: …/release.yml@<ref>` line that pins the machinery, so one pin governs
machinery and doctrine alike, and a second pin cannot fall out of sync.
**Bootstrap on adoption**: add the release and labels callers first (the pin
must exist — the mirror is verified against it), then run `--fix` once from
the repo root and commit `.ceremony/` together with the callers:
```sh
curl -fsSL "https://raw.githubusercontent.com/heavy-duty/ceremony/<pinned-tag>/actions/docs-sync/docs-sync.sh" \
| bash -s -- --fix
```
If the repo has no root `AGENTS.md`, `--fix` also scaffolds the thin stub
that routes agents to `.ceremony/AGENTS.md` — created once, never
overwritten; it is per-repo content the moment you edit it, so `--check`
asserts only that it exists.
Bumping the pin re-syncs the mirror in the same PR —
[the pin-bump procedure](#the-pin-bump-procedure).
## Version pinning
- **Pin an exact ceremony release tag** — `@0.1.0`, never a branch and
never a moving major pointer: the family pins things and reviews
updates ([#1 D2](https://github.com/heavy-duty/ceremony/issues/1)).
Every `uses:` of this repo in the consumer — the two workflow callers
and the guard steps — names the same tag.
- **Bump by PR, every reference together.** Before bumping, read the
ceremony's own `CHANGELOG.md` section for the new version (the release
body on its
[releases page](https://github.com/heavy-duty/ceremony/releases) is
that section, verbatim). One bump PR updates **every** ceremony `uses:`
reference in the repo to the new tag — the workflow callers *and* each
guard step. The exact count is tag-dependent: it is the workflow caller
or callers plus the guards that the pinned tag carries. Changing only
one line leaves the consumer split across ceremony versions, which the
same-tag rule above forbids. A repo that has adopted the agent team flow
additionally bumps the mirror in the same PR —
[the pin-bump procedure](#the-pin-bump-procedure).
- **One pin governs machinery and doctrine.** The ref in the consumer's
`release.yml` `uses:` line is the single pin: `docs-sync` reads it from
exactly there and verifies the `.ceremony/` mirror against it — there
is no second pin to fall out of sync (#19).
## The changelog rule
The portable version of the family's contributor rule — the repo's own
CONTRIBUTING may sharpen it, but this is the floor the guards assume:
- **Every PR that changes behavior writes one fragment**:
`changelog.d/<issue>.md`, named for the authorizing issue —
`<repo>-<issue>.md` for cross-repo work carrying `Part of <repo>#N` —
so the name is known at claim time and two builders can only collide by
working the same issue (#112 D2). Never an edit to `CHANGELOG.md`: the
release PR assembles the section
([below](#assembling-a-release-section)).
The sole exception is the release PR: it writes no fragment. It consumes
the directory and stamps the section, so a fragment it created would be
absent from
[`changelog-assembled`](https://github.com/heavy-duty/ceremony/blob/a602fd0/actions/changelog-assembled/changelog-assembled.sh)'s
merge-base replay if consumed, or refused by
[`changelog-armed`](https://github.com/heavy-duty/ceremony/blob/a602fd0/actions/changelog-armed/changelog-armed.sh)
if left to survive into the next release. A change that must ship inside
the release PR therefore ships without an entry. If it can wait and wants
an entry, land it as an ordinary PR before the release PR, then rebase and
re-assemble the release.
- **The fragment is the prose, not a description of it** (#112 D3): the
exact lines that will be published — no front-matter, no `## ` heading
(that one is the assembler's to write). `changelog-armed` refuses a
malformed fragment on the PR that wrote it.
- **Grouped repos group inside the fragment**: `### Added`, `### Changed`,
`### Fixed` headings with bullets under them; create `Deprecated`,
`Removed`, or `Security` only when a change genuinely needs that rarer
kind. A repo is grouped or flat, never both (#112 D4). The assembler
merges groups in canonical order — Added, Changed, Fixed, Removed,
Deprecated, Security, then anything else first-seen — and inside a
group entries read newest issue first (#112 D5). Which shape binds is
inferred from the newest published section, unless an optional sentinel
`changelog.d/shape` — one line, exactly `flat` or `grouped` — declares
it and outranks the inference (#182). To flip a repo's shape, land one
PR that adds the sentinel and converts every pending fragment to the
declared shape, bullets byte-identical; the sentinel stays after the
release, as the declaration a reader in the directory finds.
- **One line: say what changed, and stop.** Lead with the surface, not
the mechanism — "`state:needs-human` is set at handoff" beats "the
labels workflow now also wakes on `labeled`". The why and the how
belong in the PR body, where anyone chasing the reasoning already goes.
- **Cite the issue or PR** — `(#141)`.
- **Mark a breaking change** with a leading `BREAKING:`.
- A repo not yet on fragment mode — no `changelog.d/` — keeps the legacy
floor until its conversion: one line under `## Unreleased`, inserted
**above** the heading below it, never over it (replacing a shipped
heading deletes that release's section silently — box#122, why the
[monotonic guard](../README.md#changelog-monotonic--shipped-headings-are-append-only)
exists), appended under a standing `### ` heading where the repo groups.
## Assembling a release section
The ceremony PR's changelog stamp is one command, run **by hand, never in
CI** — the assembled section must land in the release PR's diff, where
the panel reads it (#112 D12). A consumer runs the tool from a ceremony
checkout at its own pin:
```sh
git clone --depth 1 --branch <pinned-tag> https://github.com/heavy-duty/ceremony /tmp/ceremony
/tmp/ceremony/bin/changelog-assemble <X.Y.Z>
```
Run it at the repo root. It folds every `changelog.d/` fragment into a
new `## X.Y.Z — DATE` section on top of `CHANGELOG.md` (DATE is today's
UTC date; pass one as a second argument to choose it) and deletes the
fragments it consumed — commit both halves together. `--check` prints the
would-be section body without touching anything; read it before running
the real thing. In CI, `changelog-assembled` replays the run from the
merge base and refuses a stamp that is not byte-for-byte what the
fragments assemble to — a mis-run hand step fails the PR, not the
published release.
## Adopting the agent team flow
The team flow (discussion → triage → issue → build → review → human
merge) is **optional per repo and separable from the release ceremony**:
a repo can adopt release-only and take the team flow later — incubator's
initial posture (#16). The model is this repo's own
[CONTRIBUTING](../CONTRIBUTING.md) ("How the other repos use this");
this is the checklist:
- [ ] **Enable Discussions** — the triage door exists or the pipeline
has no intake.
- [ ] **Vendor the doctrine**: run `docs-sync --fix` (#19) to materialize
`.ceremony/{AGENTS,TRIAGE,BUILDER,REVIEWER,LABELS}.md` —
byte-identical to this repo at the pinned ref — plus the generated
`.ceremony/README.md` (machine-managed marker) and, if the repo has
none, the thin root `AGENTS.md` stub ("governed by
heavy-duty/ceremony; read `.ceremony/AGENTS.md` first; repo
specifics in CONTRIBUTING"). The stub is scaffolded once and never
overwritten; the mirror is machine-written and never hand-edited.
Commit `.ceremony/` together with the workflow callers.
- [ ] **Guard the mirror in CI**: add the `docs-sync` check step
alongside the other guards —
```yaml
- uses: heavy-duty/ceremony/actions/docs-sync@<pinned-tag>
```
(`mode: check` is the default.) Hand-editing a vendored file, or
bumping the pin without re-syncing, goes red (#19).
- [ ] **Reduce tool-specific files** (`CLAUDE.md`, …) to one pointer line
at the root `AGENTS.md`, so every harness converges on the same
router.
- [ ] **Point CONTRIBUTING at the mirror**: a short header telling agents
to read `.ceremony/` first — agents never leave the working tree to
read the rules — followed by only what is genuinely per-repo: the
review panel roster, the `scope:*` set, the drill meaning, the
repo's code conventions.
- [ ] **Name the review panel**: the roster table in CONTRIBUTING and the
`panel=` line in `.github/labels.conf` — the required verdicts for
any PR are the panel minus its author (#10).
- [ ] **Bootstrap the issue-flow labels**: the labels
`workflow_dispatch` once ([above](#labels-automation)), or the hand
commands in [LABELS.md](../LABELS.md).
- [ ] **State the single-writer rule** in the repo's own docs: only
triage mints issues; everyone else opens discussions.
### The pin-bump procedure
Bumping the ceremony pin is **one PR carrying both halves**: every
ceremony `uses:` reference — the workflow callers *and* each guard step,
[all to the same new tag](#version-pinning) — and the re-synced
`.ceremony/` mirror —
run `docs-sync --fix` locally, or let the red `--check` on the bump PR
tell you what is stale. The CI guard is what makes a half-done bump —
pin without mirror, or mirror without pin — unmergeable (#19). This is
how a process change rolls out to a governed repo: deliberately, per
repo, reviewed.
View git blame Copy permalink