Gate the build on Docker lint and test phases (closes #40, closes #30)

Per the owner ruling on issue 40, linting and testing are phases of the
main Dockerfile rather than a separate lint file. script/lint and
script/test build one phase each by name with caching disabled, and the
final stage copies a harmless file from each so the image cannot be built
unless both passed. A stage that is not the last is built only when
something depends on it or --target names it, so the gates are invoked by
name and the edges kept. script/check runs the gates and builds no image
of its own; script/cibuild bootstraps first, because CI runs it alone and
fmt-check is native. fmt and fmt-check source nvm for the pinned node
before calling yarn, which bootstrap installs but leaves off its caller's
PATH. Every build in script/ is tagged and uncached. Issue 30 closes too:
a container has its own lint cache and lock.

Model: opus-5

.dockerignore

@@ -1,3 +1,58 @@
+# .dockerignore does NOT use .gitignore semantics. Docker matches with
+# moby/patternmatcher: filepath.Match plus `**`, so `*` does not cross
+# `/` and an unprefixed pattern is anchored at the context root. Every
+# depth-independent pattern therefore needs `**/`, or `config/.env` and
+# `certs/server.key` still ship while this file reads as solved. Only
+# genuinely root-anchored entries go unprefixed. Never transplant these
+# into .gitignore, where `**/` is wrong.
+#
+# Matching is case-sensitive, so secrets use character ranges rather
+# than an ALL-CAPS twin, which would still miss `Server.Key`.
+#
+# Extend with this repo's own host-built artifacts, written anchored:
+# `/myapp`, never `**/myapp`, which also matches `cmd/myapp/` and
+# deletes the package directory from the context.
+
+# Excluding .git means `git describe` cannot run in any build stage and
+# fails quietly there; pass the version in with --build-arg VERSION.
 .git
-node_modules
+
-.DS_Store
+# Agent scratch: one full checkout of the repo per in-flight agent.
+# Anchored because it occurs once where agents run at the repo root.
+# KNOWN GAP: a repo running agents in subdirectories still ships
+# `services/api/.claude/` and must add its own anchored entry.
+.claude
+
+# Environment files. `*.env` covers bare `.env` and the `prod.env`
+# convention. Re-include a committed template with a negation if the
+# build needs one: `!docs/example.env`.
+**/*.[eE][nN][vV]
+**/.[eE][nN][vV].*
+**/.[eE][nN][vV][rR][cC]
+
+# Private keys and the bundles carrying them. Public certificates
+# (*.crt, *.cer) are deliberately absent: they are legitimate inputs.
+**/*.[pP][eE][mM]
+**/*.[kK][eE][yY]
+**/*.[pP]12
+**/*.[pP][fF][xX]
+**/[iI][dD]_[rR][sS][aA]
+**/[iI][dD]_[dD][sS][aA]
+**/[iI][dD]_[eE][cC][dD][sS][aA]
+**/[iI][dD]_[eE][dD]25519
+
+# Dependencies: restored inside the image, never copied in.
+**/node_modules
+
+# OS metadata.
+**/.DS_Store
+**/Thumbs.db
+
+# Editor state: never a build input, and it churns COPY.
+**/*.swp
+**/*.swo
+**/*~
+**/*.bak
+**/.idea
+**/.vscode
+**/*.sublime-*

.gitignore

@@ -11,6 +11,12 @@ Thumbs.db
 .vscode/
 *.sublime-*
 
+# Agent scratch (worktrees of this repo, created and destroyed by
+# in-flight tooling). Unanchored: .gitignore patterns already match at
+# every depth, so no prefix is wanted here. This is not a .dockerignore
+# entry and must not be given a `**/` prefix on the way into one.
+.claude/
+
 # Node
 node_modules/
 

.golangci.yml

@@ -13,7 +13,6 @@ linters:
   disable:
     # Genuinely incompatible with project patterns
     - exhaustruct      # Requires all struct fields
-    - depguard         # Dependency allow/block lists
     - godot            # Requires comments to end with periods
     - wsl              # Deprecated, replaced by wsl_v5
     - wrapcheck        # Too verbose for internal packages
@@ -28,6 +27,32 @@ linters:
       max-complexity: 15
     dupl:
       threshold: 100
+    depguard:
+      # Test-support code must not be compiled into the shipped binary. A
+      # test-support package exists to hand a test privileges the program
+      # itself must never have, so a file that is not a test must not import
+      # one. Test files, and the files inside a package whose directory name
+      # ends in `test`, are where that code belongs, and are exempt.
+      #
+      # The deny list below is the one part of this file a repository is
+      # expected to extend, and the only part it may. depguard matches an
+      # import path against a list of prefixes, so it cannot be told "any path
+      # whose last segment ends in test"; a repository's own test-support
+      # packages have to be named here one at a time, by full import path,
+      # under a module path that differs from repository to repository. Add
+      # them; change nothing else.
+      rules:
+        test-support:
+          list-mode: lax
+          files:
+            - "$all"
+            - "!$test"
+            - "!**/*test/**"
+          deny:
+            - pkg: net/http/httptest
+              desc: >-
+                Test-support code belongs in test files and in packages whose
+                directory name ends in test, not in the shipped binary.
 
 issues:
   max-issues-per-linter: 0

Dockerfile

@@ -1,15 +1,58 @@
+# Lint phase. The linter is invoked directly rather than through `make
+# lint` or `script/lint`, which are themselves a docker build and would
+# recurse into a daemon that does not exist in a build step.
+#
 # node 22-alpine, 2026-02-22
-FROM node@sha256:e4bf2a82ad0a4037d28035ae71529873c069b13eb0455466ae0bc13363826e34
+FROM node@sha256:e4bf2a82ad0a4037d28035ae71529873c069b13eb0455466ae0bc13363826e34 AS lint
 
 WORKDIR /app
 
-# script/bootstrap installs all prerequisites (make via apk here; node
-# and yarn are already in the base image, so those steps are skipped).
-# Dependency manifests are copied first so the bootstrap layer is
-# cached until they change.
 COPY script/ script/
 COPY package.json yarn.lock ./
 RUN script/bootstrap
 
 COPY . .
-RUN make check
+
+RUN yarn run prettier --check '**/*.md' --tab-width 4 --prose-wrap always
+
+# Test phase, same shape and for the same reason.
+#
+# node 22-alpine, 2026-02-22
+FROM node@sha256:e4bf2a82ad0a4037d28035ae71529873c069b13eb0455466ae0bc13363826e34 AS test
+
+WORKDIR /app
+
+COPY script/ script/
+COPY package.json yarn.lock ./
+RUN script/bootstrap
+
+COPY . .
+
+RUN echo "No tests defined."
+
+# Development environment, and the last stage: a plain `docker build .`
+# names no target and so builds this one. Nothing is wanted from the two
+# phases above; the copies are what make BuildKit build them first, so
+# this image cannot be produced unless lint and test passed. A stage
+# appended after this one would drop all three out of a plain build.
+#
+# node 22-alpine, 2026-02-22
+FROM node@sha256:e4bf2a82ad0a4037d28035ae71529873c069b13eb0455466ae0bc13363826e34
+
+WORKDIR /app
+
+COPY --from=lint /app/package.json /dev/null
+COPY --from=test /app/package.json /dev/null
+
+# script/bootstrap installs all prerequisites. Manifests are copied
+# first so that layer stays cached until dependencies change.
+COPY script/ script/
+COPY package.json yarn.lock ./
+RUN script/bootstrap
+
+COPY . .
+
+# The version is computed on the host and passed in, because
+# .dockerignore excludes .git.
+ARG VERSION=dev
+LABEL org.opencontainers.image.version="${VERSION}"

README.md

@@ -116,16 +116,24 @@ alpine. We provide:
   `script/bootstrap`, then `script/install-precommit`
 - `script/projectname` — output the project name (our own extension); used by
   `script/docker` for the image tag
-- `script/test` — run the test suite (no tests defined here)
+- `script/test` — `docker build --no-cache --target test -t prompts-test .`,
-- `script/lint` — lint the markdown files with prettier
+  building the `test` phase of the `Dockerfile` (no tests defined here)
-- `script/fmt` — format all markdown files with prettier (writes)
+- `script/lint` — `docker build --no-cache --target lint -t prompts-lint .`,
-- `script/fmt-check` — check formatting (read-only)
+  building the `lint` phase, which runs prettier over the markdown files
+- `script/fmt` — format all markdown files with prettier (writes; native, not in
+  a container)
+- `script/fmt-check` — check formatting (read-only; native)
 - `script/check` — run all checks: `test`, `lint`, `fmt-check` (our own
-  extension)
+  extension); builds no image of its own
-- `script/docker` — build the Docker image, tagged via `script/projectname`
+- `script/docker` —
-  (byte-identical across repos)
+  `docker build --no-cache --build-arg VERSION="$version" -t prompts .`, the tag
-- `script/cibuild` — cd to the repo root and `docker build .` (what CI runs; the
+  coming from `script/projectname` (byte-identical across repos)
-  image build runs `script/check`)
+- `script/cibuild` — cd to the repo root, run `script/bootstrap`, run
+  `script/check`, compute `version` from `git describe`, then
+  `docker build --no-cache --build-arg VERSION="$version" -t prompts .` (what CI
+  runs; it bootstraps because CI checks out and runs this alone while
+  `script/fmt-check` is native, and the version is computed on the host because
+  `.dockerignore` excludes `.git`)
 - `script/precommit` — run by the git pre-commit hook (our own extension); calls
   `script/check`
 - `script/install-precommit` — installs the git pre-commit hook (our own

TODO.md

@@ -21,6 +21,38 @@ fmt-check, and commit.
 
 # Completed Steps
 
+- 2026-09-08: Moved linting and testing into Docker as phases of the main
+  `Dockerfile`, per the owner ruling on issue 40. `script/lint` and
+  `script/test` build one phase each by name with `--no-cache` — the same answer
+  issue 26 got, so no separate cache-busting mechanism survives — and the final
+  stage copies a harmless file from both, so the image cannot be built unless
+  they pass. This also closes issue 30: a container has its own result cache and
+  its own lock, so a lint verdict can no longer belong to another checkout. No
+  separate lint Dockerfile, and no `golangci-lint config verify` step.
+  `script/check` runs the gates and nothing else, and `script/cibuild`
+  bootstraps first, since it is all CI runs and `script/fmt-check` is native.
+- 2026-09-08: Kept in-repo agent scratch out of the Docker build context and out
+  of version control: `.claude/` is one full checkout of the repo per in-flight
+  agent, and under `COPY . .` all of it was reaching the image. Also closed the
+  consequence of excluding `.git` — `git describe` yields an empty version
+  inside a build stage without failing, so `script/docker` and `script/cibuild`
+  now compute the version on the host and pass `--build-arg VERSION`.
+- 2026-09-08: Closed the secret exposure in the canonical `.dockerignore`: a
+  local `.env`, `*.pem` or `*.key` was reaching the build context under
+  `COPY . .`, invisible to every git-based check. The patterns are now written
+  to `.dockerignore`'s own semantics — `**/`-prefixed so they hold at every
+  depth, case-folded with character ranges — and `REPO_POLICIES.md` requires
+  verifying by enumerating the image rather than by reading the file.
+- 2026-09-08: Made a pinned tool in `script/bootstrap` actually reach the host.
+  `REPO_POLICIES.md` now requires comparing the installed version against the
+  pin rather than testing `PATH` presence, and re-resolving the binary through
+  `PATH` after installing, so a version bump cannot be a silent no-op and a
+  shadowed install cannot report success.
+- 2026-09-08: Closed the false green in the canonical CI gate: `script/cibuild`
+  and `script/docker` now build with `--no-cache`, so the Dockerfile's check
+  layers cannot be served from cache on an unchanged tree, and the text claiming
+  a bare `docker build .` proves the checks ran is corrected in
+  `REPO_POLICIES.md`, both checklists and the Go styleguide.
 - 2026-08-07: Set the canonical `.golangci.yml` to the org-standard v2-schema
   config already deployed byte-identical across the org's Go repos (settings
   under `linters.settings` so thresholds like lll/funlen/cyclop/dupl actually

prompts/CODE_STYLEGUIDE_GO.md

@@ -1,6 +1,6 @@
 ---
 title: Code Styleguide — Go
-last_modified: 2026-03-18
+last_modified: 2026-09-08
 ---
 
 1.  Try to hard wrap long lines at 77 characters or less.
@@ -49,7 +49,14 @@ last_modified: 2026-03-18
     ```
 
     ```make
-    VERSION := $(shell git describe --always --dirty)
+    # ?= rather than := because this `$(shell git describe ...)` is only
+    # correct on the host: `.dockerignore` excludes `.git`, so evaluated
+    # inside a build stage it expands to the empty string without failing
+    # and the binary reports no version. The version is computed on the
+    # host by `script/docker` / `script/cibuild` and passed with
+    # `--build-arg VERSION=...`; where a build stage invokes make,
+    # `ARG VERSION` puts it in the environment and `?=` defers to it.
+    VERSION ?= $(shell git describe --always --dirty)
     BUILDARCH := $(shell uname -m)
 
     GOLDFLAGS += -X main.Version=$(VERSION)
@@ -98,12 +105,19 @@ last_modified: 2026-03-18
 
 1.  For anything beyond a simple script or tool, or anything that is going to
     run in any sort of "production" anywhere, make sure it passes
-    `golangci-lint`.
+    `golangci-lint`. Run it with `make lint`, never by invoking the binary: the
+    linter runs as a phase of the `Dockerfile` and is not installed on the host
+    by any repo. Invoked directly on a shared host it reads a result cache keyed
+    on file content rather than location, and a host-global lock, so its answer
+    may belong to another checkout entirely.
 
 1.  Write a `Dockerfile` for every repo, even if it only runs the tests and
-    linting. `docker build .` should always make sure that the code is in an
+    linting. It carries the lint and test phases, and the final stage depends on
-    able-to-be-compiled state, linted, and any tests run. The Docker build
+    both, so a build makes sure the code is in an able-to-be-compiled state,
-    should fail if linting doesn't pass.
+    linted, and its tests run. Go through `script/cibuild` or `script/docker`
+    rather than a bare `docker build .`: they pass `--no-cache`, without which
+    an unchanged tree serves the gate layers from cache and the build reports a
+    green it never ran.
 
 1.  Every repo must have a `Makefile`. See
     [Repository Policies](https://git.eeqj.de/sneak/prompts/raw/branch/main/prompts/REPO_POLICIES.md)
@@ -124,10 +138,15 @@ last_modified: 2026-03-18
 
 1.  Keep the `main()` function as small as possible.
 
-1.  Keep the `main` package as small as possible. Move as much code as is
+1.  Keep the `main` package as small as possible. Each `cmd/<name>/` directory
-    feasible to a library package, even if it's an internal one. `main` is just
+    contains a single `main.go` whose body is one call into library code (for
-    an entrypoint to your code, not a place for implementations. Exception:
+    example `os.Exit(cli.Main())` calling `internal/cli`). All CLI logic — flag
-    single-file scripts.
+    parsing, subcommand dispatch, argument handling, output formatting — lives
+    in `internal/` or `pkg/`, not in `cmd/`. `main` is just an entrypoint to
+    your code, not a place for implementations. Exception: single-file scripts.
+
+1.  No project logic outside `internal/` or `pkg/`. Anything in `cmd/` is a thin
+    entrypoint only.
 
 1.  HTTP HandleFuncs should be returned from methods or functions that need to
     handle HTTP requests. Don't use methods or your top level functions as

prompts/EXISTING_REPO_CHECKLIST.md

@@ -1,6 +1,6 @@
 ---
 title: Existing Repo Checklist
-last_modified: 2026-07-06
+last_modified: 2026-09-08
 ---
 
 Use this checklist when beginning work in a repo that may not yet conform to our
@@ -24,15 +24,46 @@ with your task.
 - [ ] `LICENSE` file exists and matches the README
 - [ ] `REPO_POLICIES.md` exists and version date is current — fetch from
       `https://git.eeqj.de/sneak/prompts/raw/branch/main/prompts/REPO_POLICIES.md`
-- [ ] `.gitignore` is comprehensive (OS, editor, language artifacts, secrets) —
+- [ ] `.gitignore` is comprehensive (OS, editor, agent scratch, language
-      fetch from `https://git.eeqj.de/sneak/prompts/raw/branch/main/.gitignore`
+      artifacts, secrets) — fetch from
-      if missing
+      `https://git.eeqj.de/sneak/prompts/raw/branch/main/.gitignore` if missing.
+      An existing repo usually has a hand-written one that is never re-fetched,
+      so check the entries rather than the file's presence.
 - [ ] `.editorconfig` exists — fetch from
       `https://git.eeqj.de/sneak/prompts/raw/branch/main/.editorconfig`
-- [ ] `Dockerfile` and `.dockerignore` exist; Dockerfile runs `make check` as a
+- [ ] `Dockerfile` and `.dockerignore` exist; the Dockerfile carries a `lint`
-      build step — fetch `.dockerignore` from
+      phase and a `test` phase, and the final stage carries a `COPY --from=` of
+      a harmless file from each — fetch `.dockerignore` from
       `https://git.eeqj.de/sneak/prompts/raw/branch/main/.dockerignore`
-- [ ] Gitea Actions workflow in `.gitea/workflows/` runs `docker build .` on
+- [ ] Nothing has been appended after the final stage, and the gate phases are
+      reachable from it. A stage nothing depends on is built only when
+      `--target` names it, so a lost `COPY --from=` edge leaves `docker build .`
+      passing while the gate never runs. Confirm by planting a violation, not by
+      reading the file.
+- [ ] The gate phases invoke their tools directly, never through `make lint` or
+      `script/test` — those are themselves a `docker build` and would recurse
+      inside a build step
+- [ ] Every depth-independent pattern in `.dockerignore` carries a `**/` prefix,
+      only genuinely root-anchored entries such as `.git` are unprefixed, and
+      `.gitignore`'s patterns have not been transplanted unmodified — the
+      transplanted form leaves `config/.env` and `certs/server.key` in the build
+      context while reading as solved
+- [ ] `.dockerignore` excludes the repo's own host-built artifacts (compiled
+      binaries, test binaries, coverage output), written root-anchored —
+      `/myapp`, never `**/myapp`. An existing repo is where such a binary is
+      likeliest to already be sitting in the build context, invisible to git.
+- [ ] `.claude/` is in `.gitignore` (unanchored) and `.claude` in
+      `.dockerignore` (anchored, no `**/` prefix). Agent worktrees are entire
+      checkouts of the repo, so they inflate the context by a multiple of it and
+      can copy another session's unreviewed work into an image layer. If agents
+      here run anywhere other than the repo root, the anchored entry misses
+      `services/api/.claude/`: add anchored entries for those directories.
+- [ ] If the repo embeds a version in a binary, that version is computed on the
+      host and passed with `--build-arg VERSION=...` by `script/docker` and
+      `script/cibuild`, and no stage calls `git describe`. A tag-derived version
+      additionally needs `fetch-depth: 0` on the CI checkout step, which clones
+      shallow and fetches no tags by default.
+- [ ] Gitea Actions workflow in `.gitea/workflows/` runs `script/cibuild` on
       push — reference
       `https://git.eeqj.de/sneak/prompts/raw/branch/main/.gitea/workflows/check.yml`
 - [ ] Language-specific config:
@@ -58,6 +89,29 @@ with your task.
       `script/install-precommit`, shimmed by `make hooks`) runs it
 - [ ] README has an **Entrypoints** section documenting the `script/`
       entrypoints and linking the standard
+- [ ] `script/lint` and `script/test` build their phase by name
+      (`docker build --no-cache --target <phase> -t <name>-<phase> .`), and no
+      host invocation anywhere in the repo can produce a lint verdict — grep for
+      the linter's own name across `script/`, the `Makefile` and CI config, not
+      just `script/lint`. A second path is likeliest here: a `make lint-fast`,
+      an older host-versus-container branch, or a CI step calling the binary
+      directly. `script/fmt` and `script/fmt-check` are expected hits and stay
+      on the host.
+- [ ] Every `docker build` in `script/` is tagged — an untagged one leaves a
+      dangling image behind on every run, on every host and CI runner
+- [ ] `script/cibuild` runs `script/bootstrap` before `script/check`, and builds
+      the image with `--no-cache`. Without the bootstrap the CI run dies in
+      `script/fmt-check`, which runs the formatter on the host and finds nothing
+      installed.
+- [ ] `script/fmt` and `script/fmt-check` source nvm for the pinned node version
+      before invoking `yarn`, as `script/bootstrap`'s own install step does.
+      `script/bootstrap` leaves the node and yarn it installs off the `PATH` of
+      the shell that called it, so a bare `yarn` exits 127 on a runner carrying
+      nothing but docker and git.
+- [ ] `script/bootstrap` installs no linter of its own — delete the block, its
+      version variables and its call site. A JS repo's `yarn install` stays; it
+      brings a linter along with every other dependency, and no verdict is taken
+      from it.
 - [ ] `make check` does not modify any files in the repo
 - [ ] `make test` has a 90-second timeout and completes within the 60-second
       hard cap (over 20 seconds is green but must be filed as an improvement
@@ -106,5 +160,10 @@ with your task.
 # Final
 
 - [ ] `make check` passes
-- [ ] `docker build` succeeds
+- [ ] `script/cibuild` succeeds in a fresh clone on a host carrying nothing but
+      docker and git, with no node or yarn on `PATH`, which is what CI has, and
+      demonstrably executed the checks — a sub-second build, or `CACHED` on a
+      gate layer, means nothing ran
+- [ ] A planted lint violation fails both `make lint` and a plain
+      `docker build .`; revert it afterwards
 - [ ] Commit and merge fixes before starting your actual task

prompts/GO_HTTP_SERVER_CONVENTIONS.md

@@ -1,6 +1,6 @@
 ---
 title: Go HTTP Server Conventions
-last_modified: 2026-02-22
+last_modified: 2026-09-08
 ---
 
 This document defines the architectural patterns, design decisions, and
@@ -991,7 +991,14 @@ func main() {
 Use ldflags to inject version information at build time:
 
 ```makefile
-VERSION := $(shell git describe --tags --always)
+# ?= rather than := because this `$(shell git describe ...)` is only correct
+# on the host: `.dockerignore` excludes `.git`, so evaluated inside a build
+# stage it expands to the empty string without failing and the binary reports
+# no version. The version is computed on the host by `script/docker` /
+# `script/cibuild` and passed with `--build-arg VERSION=...`; where the build
+# stage invokes make, `ARG VERSION` puts it in the environment and `?=` defers
+# to it.
+VERSION ?= $(shell git describe --tags --always)
 BUILDARCH := $(shell go env GOARCH)
 
 build:

prompts/NEW_REPO_CHECKLIST.md

@@ -1,6 +1,6 @@
 ---
 title: New Repo Checklist
-last_modified: 2026-07-06
+last_modified: 2026-09-08
 ---
 
 Use this checklist when creating a new repository from scratch. Follow the steps
@@ -35,7 +35,11 @@ Template files can be fetched from:
 
 - [ ] `.gitignore` — fetch from
       `https://git.eeqj.de/sneak/prompts/raw/branch/main/.gitignore`, extend for
-      language-specific artifacts
+      language-specific artifacts. Extensions are written to `.gitignore`'s own
+      semantics, where an unanchored pattern already matches at every depth:
+      never add a `**/` prefix here, which is a `.dockerignore` form. The
+      canonical file already carries `.claude/` so agent worktrees cannot be
+      committed by accident.
 - [ ] `.editorconfig` — fetch from
       `https://git.eeqj.de/sneak/prompts/raw/branch/main/.editorconfig`
 - [ ] `Makefile` — fetch from
@@ -52,9 +56,29 @@ Template files can be fetched from:
       `https://git.eeqj.de/sneak/prompts/raw/branch/main/prompts/REPO_POLICIES.md`
 - [ ] `Dockerfile` and `.dockerignore` — fetch `.dockerignore` from
       `https://git.eeqj.de/sneak/prompts/raw/branch/main/.dockerignore`
-    - All Dockerfiles must run `make check` as a build step
+    - Extend `.dockerignore` with the repo's own host-built artifacts, giving
-    - Server: also builds and runs the application
+      every depth-independent pattern a `**/` prefix — but write a repo-root
-    - Non-server: brings up dev environment and runs `make check`
+      binary anchored, `/myapp` and never `**/myapp`, which would also match
+      `cmd/myapp/` and delete the package directory. Do not transplant
+      `.gitignore`'s patterns: `.dockerignore` anchors an unprefixed pattern at
+      the context root, so the copied form leaves `config/.env` in the build
+      context while reading as solved. The canonical file's `.claude` entry is
+      anchored for the same reason as a repo-root binary; leave it that way, but
+      note that it only covers agents running at the repo root — if this repo
+      will run them in subdirectories, `services/api/.claude/` needs its own
+      anchored entry.
+    - If the image embeds a version in a binary, the version is computed on the
+      host and passed with `--build-arg VERSION=...`, and `ARG VERSION=dev` is
+      declared in the stage that compiles. **No stage calls `git describe`** —
+      `.dockerignore` excludes `.git`, so it yields an empty version without
+      failing the build.
+    - The Dockerfile carries a `lint` phase and a `test` phase, each invoking
+      its tool directly rather than through `make` or `script/`, and the final
+      stage carries a `COPY --from=` of a harmless file from each so the image
+      cannot be built unless both passed. Keep the final stage last: a stage
+      nothing depends on is built only when `--target` names it.
+    - Server: the final stage builds and runs the application
+    - Non-server: the final stage brings up the dev environment
     - Image pinned by sha256 hash with version/date comment
 - [ ] Gitea Actions workflow at `.gitea/workflows/check.yml` that runs
       `script/cibuild` on push — reference
@@ -80,18 +104,34 @@ are thin shims calling them. Model scripts:
       installs
 - [ ] `script/setup` / `make setup` — readies a fresh clone: runs `bootstrap`,
       then `install-precommit`, plus repo-specific init
-- [ ] `script/test` / `make test` — runs real tests, not a no-op (90-second
+- [ ] `script/test` / `make test` — `docker build --no-cache --target test .`,
-      timeout, 60-second hard cap on wall time)
+      tagged; the phase runs real tests, not a no-op (90-second timeout,
-- [ ] `script/lint` / `make lint` — runs linter
+      60-second hard cap on wall time)
-- [ ] `script/fmt` / `make fmt` — formats code (writes)
+- [ ] `script/lint` / `make lint` — `docker build --no-cache --target lint .`,
-- [ ] `script/fmt-check` / `make fmt-check` — checks formatting (read-only)
+      tagged. No lint verdict may come from a host invocation of the linter.
+- [ ] `script/fmt` / `make fmt` — formats code (writes; native, never in a
+      container)
+- [ ] `script/fmt-check` / `make fmt-check` — checks formatting (read-only;
+      native)
 - [ ] `script/check` / `make check` — runs `test`, `lint`, `fmt-check`; must not
       modify files
 - [ ] `script/projectname` — outputs the project name (used by `script/docker`
       for the image tag)
 - [ ] `script/docker` / `make docker` — builds Docker image, tagged via
-      `script/projectname` (byte-identical across repos)
+      `script/projectname` (byte-identical across repos); `--no-cache`, plus the
-- [ ] `script/cibuild` — cd to repo root, `docker build .` (what CI runs)
+      version as a build arg
+- [ ] `script/cibuild` — cd to repo root, run `script/bootstrap`, run
+      `script/check`, then
+      `docker build --no-cache --build-arg VERSION="$version" .` (what CI runs).
+      The bootstrap is required: CI checks out and runs this alone, and
+      `script/fmt-check` runs the formatter on the host.
+- [ ] `script/fmt` and `script/fmt-check` source nvm for the pinned node version
+      before invoking `yarn`, as `script/bootstrap`'s own install step does.
+      `script/bootstrap` leaves the node and yarn it installs off the `PATH` of
+      the shell that called it, so a bare `yarn` exits 127 on a runner carrying
+      nothing but docker and git.
+- [ ] Every `docker build` in `script/` is tagged, so no invocation leaves a
+      dangling image behind
 - [ ] `script/precommit` — called by the pre-commit hook; runs `script/check`
 - [ ] `script/install-precommit` — installs the pre-commit hook that runs
       `script/precommit`
@@ -103,7 +143,15 @@ are thin shims calling them. Model scripts:
 
 - [ ] `make check` passes
 - [ ] `make docker` succeeds
-- [ ] No secrets in repo
+- [ ] `script/cibuild` succeeds in a fresh clone on a host carrying nothing but
+      docker and git, with no node or yarn on `PATH`, which is what CI has, and
+      demonstrably executed the checks — a sub-second build, or `CACHED` on a
+      gate layer, means nothing ran
+- [ ] Plant a lint violation and confirm both `make lint` and a plain
+      `docker build .` fail on it; revert. A plain build that passes proves the
+      final stage is missing its `COPY --from=` edge to the gate phases.
+- [ ] No secrets in repo, and none in the build context: enumerate a probe image
+      rather than reading `.dockerignore`
 - [ ] No mutable image/package references
 - [ ] No unnecessary files in repo root
 - [ ] All dates written as YYYY-MM-DD

prompts/REPO_POLICIES.md

@@ -1,6 +1,6 @@
 ---
 title: Repository Policies
-last_modified: 2026-08-19
+last_modified: 2026-09-08
 ---
 
 This document covers repository structure, tooling, and workflow standards. Code
@@ -60,17 +60,28 @@ style conventions are in separate documents:
   prerequisite since nvm requires bash. yarn is then pinned via
   `corepack prepare yarn@<version> --activate`. Never install "latest" or "lts";
   always exact versions. `script/cibuild` runs the CI build: it changes to the
-  repo root and runs `docker build .`; the Gitea workflow calls it. Four further
+  repo root, runs `script/bootstrap`, runs `script/check`, and builds the image
-  scripts are our own extensions to the standard: `script/check` runs
+  with the version; the Gitea workflow calls it. **`script/cibuild` runs
-  `script/test`, `script/lint`, and `script/fmt-check`; `script/precommit` is
+  `script/bootstrap` first**, because the workflow checks out the repo and runs
-  what the git pre-commit hook runs, and it calls `script/check`;
+  nothing else, while `script/fmt-check` runs the formatter on the host: on a
-  `script/install-precommit` installs the git pre-commit hook (the `make hooks`
+  pristine checkout with nothing installed the run dies there, after the
-  target shims to it); and `script/projectname` (literally that filename) simply
+  containerised gates have passed. **The bootstrap alone is not enough**:
-  outputs the project's name. Scripts that need the name call
+  `script/bootstrap` installs node and yarn under nvm and leaves neither on the
-  `script/projectname` — e.g. `script/docker` assembles its image tag from it —
+  `PATH` of the shell that called it, so a bare `yarn` still exits 127. The host
-  so those scripts stay byte-identical across all repos. Repo-type-specific
+  entrypoints that need yarn — `script/fmt` and `script/fmt-check` — therefore
-  pre-commit extras (e.g. `go mod tidy` verification in Go repos) belong in
+  source nvm for the pinned node version before invoking it, exactly as
-  `script/precommit`, not in the hook itself. Model scripts are at
+  `script/bootstrap`'s own install step does. A runner carrying nothing but
+  docker and git then gets through `script/check`. Four further scripts are our
+  own extensions to the standard: `script/check` runs `script/test`,
+  `script/lint` and `script/fmt-check`; `script/precommit` is what the git
+  pre-commit hook runs, and it calls `script/check`; `script/install-precommit`
+  installs the git pre-commit hook (the `make hooks` target shims to it); and
+  `script/projectname` (literally that filename) simply outputs the project's
+  name. Scripts that need the name call `script/projectname` — e.g.
+  `script/docker` assembles its image tag from it — so those scripts stay
+  byte-identical across all repos. Repo-type-specific pre-commit extras (e.g.
+  `go mod tidy` verification in Go repos) belong in `script/precommit`, not in
+  the hook itself. Model scripts are at
   `https://git.eeqj.de/sneak/prompts/raw/branch/main/script/<name>`. The README
   must document the provided scripts in an **Entrypoints** section (see the
   README requirements below).
@@ -89,87 +100,140 @@ style conventions are in separate documents:
   contributor should be able to understand the entire development workflow by
   reading the Makefile.
 
-- Every repo should have a `Dockerfile`. All Dockerfiles must run `make check`
+- Every repo should have a `Dockerfile`, and it carries the repo's gates: a
-  as a build step so the build fails if the branch is not green. For non-server
+  `lint` phase and a `test` phase, with the final stage depending on both so the
-  repos, the Dockerfile should bring up a development environment and run
+  image cannot be built unless they pass. For non-server repos the final stage
-  `make check`. For server repos, `make check` should run as an early build
+  brings up a development environment; for server repos it is the runtime image.
-  stage before the final image is assembled. Dockerfiles install development
+  Dockerfiles install development prerequisites by running `script/bootstrap`
-  prerequisites by running `script/bootstrap` rather than duplicating installs
+  rather than duplicating installs inline; COPY `script/` and the dependency
-  inline; COPY `script/` and the dependency manifests (`package.json` +
+  manifests (`package.json` + `yarn.lock`, `go.mod` + `go.sum`, etc.) before
-  `yarn.lock`, `go.mod` + `go.sum`, etc.) before running it so the bootstrap
+  running it.
-  layer stays cached until dependencies change.
 
-- **Dockerfiles must use a separate lint stage for fail-fast feedback.** Go
+- **Linting and testing run in Docker, as phases of the `Dockerfile`.** There is
-  repos use a multistage build where linting runs in an independent stage based
+  no separate lint file. `script/lint` and `script/test` each build one phase
-  on the `golangci/golangci-lint` image (pinned by hash). This stage runs
+  and nothing else:
-  `make fmt-check` and `make lint` before the full build begins. The build stage
-  then declares an explicit dependency on the lint stage via
-  `COPY --from=lint /src/go.sum /dev/null`, which forces BuildKit to complete
-  linting before proceeding to compilation and tests. This ensures lint failures
-  surface in seconds rather than minutes, without blocking on dependency
-  download or compilation in the build stage.
 
-    The standard pattern for a Go repo Dockerfile is:
+    ```sh
+    docker build --no-cache --target lint -t "$(script/projectname)-lint" .
+    docker build --no-cache --target test -t "$(script/projectname)-test" .
+    ```
+
+    **A stage that is not the last one in the file is built only when the final
+    stage's chain depends on it, or when `--target` names it.** That is why the
+    two gates are always invoked by name here, and why the final stage carries a
+    `COPY --from=` of a harmless file from each of them: without that edge a
+    plain `docker build .` builds the last stage alone and exits 0 having linted
+    and tested nothing.
+
+    **Every `docker build` in `script/` is tagged**, here and in
+    `script/cibuild` and `script/docker`. An untagged build leaves a dangling
+    image behind on every invocation, on every developer host and every CI
+    runner; a tagged one replaces the previous image.
+
+    Inside a phase the tool is invoked directly — `golangci-lint`, `go test`,
+    `eslint`, `prettier` — never through `make lint` or `script/test`, which are
+    themselves a `docker build` and would recurse into a daemon that does not
+    exist in a build step. Formatting is the exception and stays on the host:
+    `script/fmt` writes the working tree, and `script/fmt-check` is its
+    read-only twin.
+
+    **No lint verdict may come from a host invocation of the linter.** On a
+    shared host golangci-lint reads a result cache keyed on file content rather
+    than location, so a second checkout of the same content is served the first
+    one's findings, and a host-global lock in `$TMPDIR` makes concurrent runs
+    exit non-zero with `parallel golangci-lint is running` — a status a caller
+    cannot tell from real findings. Both have produced wrong verdicts in this
+    org, in both directions. A container has its own cache, its own `TMPDIR` and
+    a digest-pinned binary, so neither is reachable.
+
+- **Any build that runs checks is built with `--no-cache`.** Docker invalidates
+  a `COPY` layer only when the copied content changes, so on an unchanged tree
+  the check `RUN` is served from cache, nothing executes, and the build still
+  exits 0. Every `docker build` in `script/` therefore passes `--no-cache`:
+  `script/lint`, `script/test`, `script/cibuild` and `script/docker` are the
+  four, and there is no fifth — `script/check` runs the two gate phases and
+  `script/fmt-check`, and builds no image of its own. A bare `docker build .` is
+  not evidence that anything ran: a sub-second build reporting success is a
+  cache hit, not a result. Never invalidate by pruning — `docker builder prune`
+  and friends destroy a build cache shared with every other build on the host.
+
+- **The gate phases are separate stages, and the build stage depends on both.**
+  The lint phase is based on the `golangci/golangci-lint` image (pinned by
+  hash), so lint failures surface in seconds rather than after a full compile,
+  and the test phase is based on the Go image. The canonical Go repo
+  `Dockerfile`:
 
     ```dockerfile
-    # Lint stage — fast feedback on formatting and lint issues
+    # Lint phase
     # golangci/golangci-lint:v2.x.x, YYYY-MM-DD
     FROM golangci/golangci-lint@sha256:... AS lint
     WORKDIR /src
     COPY go.mod go.sum ./
     RUN go mod download
     COPY . .
-    RUN make fmt-check
+    RUN golangci-lint run --config .golangci.yml ./...
-    RUN make lint
 
-    # Build stage
+    # Test phase
     # golang:1.x-alpine, YYYY-MM-DD
-    FROM golang@sha256:... AS builder
+    FROM golang@sha256:... AS test
+    WORKDIR /src
+    COPY go.mod go.sum ./
+    RUN go mod download
+    COPY . .
+    RUN go test -timeout 90s -race -cover ./... || \
+        { echo "--- Rerunning with -v for details ---"; \
+          go test -timeout 90s -race -v ./...; exit 1; }
+
+    # Build stage. Nothing is wanted from either phase above; the copies
+    # are what make BuildKit build them first, so this stage cannot run
+    # unless lint and test passed.
+    # golang:1.x-alpine, YYYY-MM-DD
+    FROM golang@sha256:... AS builder
+    COPY --from=lint /src/go.sum /dev/null
+    COPY --from=test /src/go.sum /dev/null
     WORKDIR /src
-
-    # Force BuildKit to run the lint stage before proceeding
-    COPY --from=lint /src/go.sum /dev/null
-
     COPY go.mod go.sum ./
     RUN go mod download
     COPY . .
-    RUN make test
 
     ARG VERSION=dev
     RUN CGO_ENABLED=0 go build -trimpath \
         -ldflags="-s -w -X main.Version=${VERSION}" \
         -o /app ./cmd/app/
 
-    # Runtime stage
+    # Runtime stage, and the last one
     FROM alpine@sha256:...
     COPY --from=builder /app /usr/local/bin/app
     ENTRYPOINT ["app"]
     ```
 
     Key points:
-    - The lint stage uses the `golangci/golangci-lint` image directly (it
+    - The lint phase uses the `golangci/golangci-lint` image directly (it has
-      includes both Go and the linter), so there is no need to install the
+      both Go and the linter), so nothing needs installing.
-      linter separately.
+    - `COPY --from=<phase> /src/go.sum /dev/null` is a no-op copy whose only
-    - `COPY --from=lint /src/go.sum /dev/null` is a no-op file copy that creates
+      purpose is the ordering edge. BuildKit runs stages in parallel by default,
-      a stage dependency. BuildKit runs stages in parallel by default; without
+      and a stage nothing depends on is not built at all, so without these two
-      this line, the build stage would not wait for lint to finish and a lint
+      lines a red gate would not fail the build.
-      failure might not fail the overall build.
+    - Keep the runtime stage last, and if you add a stage after it, give it the
+      same two copies. A plain `docker build .` builds the last stage's chain
+      and nothing else.
     - If the project uses `//go:embed` directives that reference build artifacts
-      (e.g. a web frontend compiled in a separate stage), the lint stage must
+      (e.g. a web frontend compiled in a separate stage), the lint phase must
       create placeholder files so the embed directives resolve. Example:
       `RUN mkdir -p web/dist && touch web/dist/index.html web/dist/style.css`.
-      The lint stage should not depend on the actual build output — it exists to
-      fail fast.
     - If the project requires CGO or system libraries for linting (e.g.
-      `vips-dev`), install them in the lint stage with `apk add`.
+      `vips-dev`), install them in the lint phase with `apk add`.
-    - The build stage runs `make test` after compilation setup. Tests run in the
+    - `ARG VERSION=dev` is declared in the stage that compiles and supplied by
-      build stage, not the lint stage, because they may require compiled
+      `script/docker` and `script/cibuild`; no stage may call `git describe`.
-      artifacts or heavier dependencies.
 
 - Every repo should have a Gitea Actions workflow (`.gitea/workflows/`) that
-  runs `script/cibuild` (which runs `docker build .`) on push. Since the
+  runs `script/cibuild` on push, and checks out the repo as its only other step.
-  Dockerfile already runs `make check`, a successful build implies all checks
+  That script bootstraps, runs the gate phases, and then builds the image, so a
-  pass.
+  successful run means every check passed; a bare `docker build .` does not
+  carry the same guarantee, because its gate phases may come from the cache. The
+  image build is uncached and so runs the gate phases a second time. That is the
+  price of the rule above, and it is worth paying: the image that ships is built
+  from a run of its own gates rather than from a cache entry.
 
 - Use platform-standard formatters: `black` for Python, `prettier` for
   JS/CSS/Markdown/HTML, `go fmt` for Go. Always use default configuration with
@@ -193,15 +257,17 @@ style conventions are in separate documents:
   suite that exceeds it fails. Under 20 seconds is the target. A suite between
   20 and 60 seconds is still green, but the overage must be filed as an
   improvement bug against that repo. Add a 90-second timeout to the test
-  invocation in the Makefile (`go test -timeout 90s`). The backstop deliberately
+  invocation (`go test -timeout 90s`). The backstop deliberately sits above the
-  sits above the hard cap so that it catches a genuinely hung test rather than a
+  hard cap so that it catches a genuinely hung test rather than a merely slow
-  merely slow one.
+  one.
 
-- **`make test` should use the conditional verbose rerun pattern.** Run tests
+- **The test command should use the conditional verbose rerun pattern.** Run
-  without `-v` (verbose) first. If tests fail, automatically rerun with `-v` to
+  tests without `-v` (verbose) first. If tests fail, automatically rerun with
-  show full output. This keeps CI logs and `docker build` output clean on
+  `-v` to show full output. This keeps CI logs and `docker build` output clean
-  success (just package/suite summaries) while providing full diagnostic detail
+  on success (just package/suite summaries) while providing full diagnostic
-  on failure (every test case, every assertion). The general shell pattern:
+  detail on failure (every test case, every assertion). The command lives in the
+  `test` phase of the `Dockerfile`, since `script/test` builds that phase; the
+  Makefile form below is the same pattern for any repo-local invocation:
 
     ```makefile
     test:
@@ -244,10 +310,83 @@ style conventions are in separate documents:
   must be in `.gitignore`. No exceptions.
 
 - `.gitignore` should be comprehensive from the start: OS files (`.DS_Store`),
-  editor files (`.swp`, `*~`), language build artifacts, and `node_modules/`.
+  editor files (`.swp`, `*~`), in-repo agent scratch directories (`.claude/`),
-  Fetch the standard `.gitignore` from
+  language build artifacts, and `node_modules/`. Fetch the standard `.gitignore`
-  `https://git.eeqj.de/sneak/prompts/raw/branch/main/.gitignore` when setting up
+  from `https://git.eeqj.de/sneak/prompts/raw/branch/main/.gitignore` when
-  a new repo.
+  setting up a new repo. These patterns are written to `.gitignore`'s own
+  semantics, in which an unanchored pattern already matches at every depth; they
+  are not a `.dockerignore` and must not be transplanted into one unmodified.
+
+- **`.dockerignore` does not use `.gitignore` semantics, and copying patterns
+  across unmodified leaves secrets in the build context.** Docker matches with
+  `moby/patternmatcher`: `filepath.Match` semantics plus a `**` extension, so
+  `*` does not cross `/` and a pattern without a leading `**/` is anchored at
+  the build-context root. A `.dockerignore` listing `.env`, `*.pem` and `*.key`
+  therefore excludes only the copies at the repository root, while `config/.env`
+  and `certs/server.key` still reach the context and can land in an image layer
+  — which is more dangerous than a short file with no secret patterns at all,
+  because it reads as solved and stops anyone looking. Give every
+  depth-independent pattern the `**/` prefix and leave only genuinely
+  root-anchored entries unprefixed: `.git`, and the repo's own host-built
+  binary, written `/myapp` and never `**/myapp`, which would also match
+  `cmd/myapp/` and delete the package directory from the context. Matching is
+  case-sensitive, and an ALL-CAPS twin per pattern still misses `Server.Key`, so
+  secret names use character ranges — `**/*.[kK][eE][yY]`, `**/*.[pP][eE][mM]`,
+  and likewise for `.envrc` and the extensionless SSH keys. Where such a pattern
+  also catches something the build needs, re-include it with a negation
+  (`!docs/example.env`); deleting the pattern reopens the exposure for every
+  other file it covers. Fetch the standard `.dockerignore` from
+  `https://git.eeqj.de/sneak/prompts/raw/branch/main/.dockerignore` and extend
+  it with the repo's own artifacts.
+
+- **In-repo agent scratch belongs in both files, written to each file's own
+  semantics.** `.claude/` holds one worktree per in-flight agent — an entire
+  additional checkout of the repo — so under `COPY . .` the build context
+  inflates by a multiple of the repo and another session's unreviewed work can
+  be copied into an image layer. In `.gitignore` the entry is `.claude/`,
+  unanchored. In `.dockerignore` it is `.claude`, anchored and with **no** `**/`
+  prefix, because the prefixed form would also delete any nested directory of
+  that name from the build. Anchoring carries a known gap that the canonical
+  `.dockerignore` states in its own comment, since consuming repos receive the
+  file and not the tracker: the directory is created in the agent's working
+  directory, so a repo running agents in subdirectories still ships
+  `services/api/.claude/` and must add its own anchored entry there.
+
+- **Excluding `.git` means `git describe` cannot run inside any build stage, and
+  it fails quietly there.** In a build stage there is no repository, so
+  `git describe` writes nothing to stdout, `-X main.Version=` comes out empty,
+  the binary reports no version at all, and the build still exits 0. Compute the
+  version on the host and thread it in as a build arg. `script/docker` and
+  `script/cibuild` do this, byte-identically across repos:
+
+    ```sh
+    # Own line: a failing command substitution inside an argument does not
+    # trip `set -e`, so the inline form degrades to an empty constant.
+    version="$(git describe --tags --always --dirty 2>/dev/null || true)"
+    [ -n "$version" ] || version="unknown"
+    docker build --no-cache \
+        --build-arg VERSION="$version" \
+        -t "$(script/projectname)" .
+    ```
+
+    `--always` makes an untagged repo yield an abbreviated commit hash rather
+    than failing, and the `[ -n "$version" ]` line is the single place the
+    fallback is applied — a live check that fires on a build from an export with
+    no `.git` and on a repository with no commits yet. Do not fold it into the
+    substitution as `|| echo unknown`, which makes the guard unreachable. The
+    Dockerfile's side is `ARG VERSION=dev` in the stage that compiles, declared
+    there because `ARG` is stage-scoped; passing `VERSION` to a repo whose
+    Dockerfile declares no such `ARG` is ignored and costs nothing, which is why
+    the scripts stay byte-identical. One consequence for CI: the standard
+    checkout action clones shallow and fetches no tags, so a repo that embeds a
+    tag-derived version must set `fetch-depth: 0` on its checkout step.
+
+- **Verify `.dockerignore` by enumerating the image, not by reading the
+  patterns.** Plant files at the root _and_ at least two directories deep, build
+  a probe image that does `COPY . .`, and list what actually landed
+  (`docker run --rm --entrypoint find IMAGE /app`). The `transferring context`
+  size is not a substitute: a nested secret is a few bytes, and BuildKit
+  transfers only the delta from the previous build.
 
 - **No build artifacts in version control.** Code-derived data (compiled
   bundles, minified output, generated assets) must never be committed to the
@@ -269,9 +408,39 @@ style conventions are in separate documents:
   byte-identical, so that no repo can quietly loosen its own linting. Linter
   configuration changes are made to the canonical copy in the `prompts` repo and
   reach consuming repos by re-vendoring; an agent may open a PR against
-  canonical, which only the user merges. The canonical golangci-lint version is
+  canonical, which only the user merges. One list is exempt from byte-identity,
-  v2.12.2 (released 2026-05-06), installed commit-pinned via
+  because it cannot be written once for every repo: the `deny` list of the
-  `go install github.com/golangci/golangci-lint/v2/cmd/golangci-lint@c0d3ddc9cf3faa61a4e378e879ece580256d76e5`.
+  `test-support` depguard rule, where a repo names its own test-support packages
+  by full import path. A repo adds entries there and changes nothing else, and a
+  re-vendor carries its entries forward. The canonical golangci-lint version is
+  v2.12.2 (released 2026-05-06), pinned as the digest of the lint phase's base
+  image
+  (`golangci/golangci-lint@sha256:5cceeef04e53efe1470638d4b4b4f5ceefd574955ab3941b2d9a68a8c9ad5240`,
+  which reports `2.12.2 built with go1.26.2 from c0d3ddc9`). That digest is the
+  only pin, since no repo installs golangci-lint on the host: bumping the
+  version means changing it and nothing else.
+
+- **`script/bootstrap` installs a pinned tool by comparing versions, never by
+  testing presence.** An `if ! command -v <tool>; then install; fi` guard tests
+  `PATH` only, so on an already-provisioned machine the pin is inert and a
+  version bump is a silent no-op — while the Dockerfile, installing into a clean
+  image, gets the pinned version, so a local `make check` and `make docker` can
+  disagree about what the tool even is. The canonical form:
+    - compares the installed version against the pin over the **whole** version
+      token; a parser that stops at the first `-` reports `2.12.2` for a host
+      running `2.12.2-rc1` and skips the install;
+    - treats absent, non-zero, empty or unrecognised `--version` output as a
+      mismatch, so the failure direction is a redundant install and never a
+      skipped one;
+    - after installing, re-resolves the binary the way callers do — `hash -r`,
+      then through `PATH`, not through the directory the installer wrote to —
+      and fails naming the resolved path, since an install that a shadowing
+      binary hides succeeds while changing nothing any caller sees;
+    - is actually called, and prints the version on both success paths: a
+      function defined and never invoked has the same exit status and the same
+      empty output as one that worked.
+
+    Keep it POSIX sh: no arrays, no `[[`, no `grep -P`.
 
 - When pinning images or packages by hash, add a comment above the reference
   with the version and date (YYYY-MM-DD).
@@ -390,7 +559,9 @@ style conventions are in separate documents:
   language-specific config). Everything else goes in a subdirectory. Canonical
   subdirectory names:
     - `bin/` — executable scripts and tools
-    - `cmd/` — Go command entrypoints
+    - `cmd/` — Go command entrypoints; thin only: one `main.go` per binary whose
+      body is a single call into `internal/` or `pkg/`, no project logic in
+      `cmd/`
     - `configs/` — configuration templates and examples
     - `deploy/` — deployment manifests (k8s, compose, terraform)
     - `docs/` — documentation and markdown (README.md stays in root)

script/check

@@ -1,6 +1,8 @@
 #!/bin/sh
 # script/check: run all checks (test, lint, fmt-check). Our own
-# extension to scripts-to-rule-them-all. Must not modify any files.
+# extension to scripts-to-rule-them-all. test and lint are Docker
+# phases; fmt-check is native, because a formatter writes the working
+# tree. Must not modify any files.
 set -eu
 
 SCRIPT_DIR="$(cd "$(dirname "$0")" && pwd -P)"

script/cibuild

@@ -1,13 +1,29 @@
 #!/bin/sh
-# script/cibuild: run the CI build. The Dockerfile runs script/check, so
+# script/cibuild: run the CI build. It bootstraps first: a CI runner
-# a successful build implies all checks pass.
+# checks out and runs this and nothing else, and script/fmt-check runs
+# the formatter on the host, which a pristine checkout cannot do.
+# --no-cache for the same reason as script/docker: the gate phases the
+# final stage depends on are RUN steps, and a cached one is a check that
+# did not run.
 set -eu
 
-ROOT="$(cd "$(dirname "$0")/.." && pwd -P)"
+SCRIPT_DIR="$(cd "$(dirname "$0")" && pwd -P)"
+ROOT="$(cd "$SCRIPT_DIR/.." && pwd -P)"
 
 main() {
     cd "$ROOT"
-    docker build .
+    "$SCRIPT_DIR/bootstrap"
+    "$SCRIPT_DIR/check"
+    # Own line: a failing command substitution inside an argument does
+    # not trip `set -e`, so the inline form degrades silently to an
+    # empty constant. VERSION is computed here because .dockerignore
+    # excludes .git, so `git describe` in a build stage yields an empty
+    # version without failing.
+    version="$(git describe --tags --always --dirty 2>/dev/null || true)"
+    [ -n "$version" ] || version="unknown"
+    docker build --no-cache \
+        --build-arg VERSION="$version" \
+        -t "$("$SCRIPT_DIR/projectname")" .
 }
 
 main "$@"

script/docker

@@ -1,6 +1,8 @@
 #!/bin/sh
 # script/docker: build the Docker image tagged with the project name.
 # Identical in all repos; the tag comes from script/projectname.
+# --no-cache because the gate phases the final stage depends on are RUN
+# steps, and a cached one is a check that did not run.
 set -eu
 
 SCRIPT_DIR="$(cd "$(dirname "$0")" && pwd -P)"
@@ -8,7 +10,16 @@ ROOT="$(cd "$SCRIPT_DIR/.." && pwd -P)"
 
 main() {
     cd "$ROOT"
-    docker build -t "$("$SCRIPT_DIR/projectname")" .
+    # Own line: a failing command substitution inside an argument does
+    # not trip `set -e`, so the inline form degrades silently to an
+    # empty constant. VERSION is computed here because .dockerignore
+    # excludes .git, so `git describe` in a build stage yields an empty
+    # version without failing.
+    version="$(git describe --tags --always --dirty 2>/dev/null || true)"
+    [ -n "$version" ] || version="unknown"
+    docker build --no-cache \
+        --build-arg VERSION="$version" \
+        -t "$("$SCRIPT_DIR/projectname")" .
 }
 
 main "$@"

script/fmt

@@ -4,9 +4,28 @@ set -eu
 
 ROOT="$(cd "$(dirname "$0")/.." && pwd -P)"
 
+# Must match the pin in script/bootstrap.
+NODE_VERSION="22.17.0"
+
+# script/bootstrap installs node and yarn under nvm and leaves neither
+# on the PATH of the shell that called it, so resolve the pinned
+# toolchain here the way bootstrap's own install step does. nvm is a
+# bash script, hence the subshell.
+run_yarn() {
+    if command -v yarn >/dev/null 2>&1; then
+        exec yarn "$@"
+    fi
+    if [ ! -s "$HOME/.nvm/nvm.sh" ]; then
+        echo "fmt: no yarn; run script/bootstrap first" >&2
+        exit 1
+    fi
+    exec bash -c '. "$HOME/.nvm/nvm.sh" && nvm use "$1" >/dev/null &&
+        shift && exec yarn "$@"' bash "$NODE_VERSION" "$@"
+}
+
 main() {
     cd "$ROOT"
-    yarn run prettier --write '**/*.md' --tab-width 4 --prose-wrap always
+    run_yarn run prettier --write '**/*.md' --tab-width 4 --prose-wrap always
 }
 
 main "$@"

script/fmt-check

@@ -4,9 +4,28 @@ set -eu
 
 ROOT="$(cd "$(dirname "$0")/.." && pwd -P)"
 
+# Must match the pin in script/bootstrap.
+NODE_VERSION="22.17.0"
+
+# script/bootstrap installs node and yarn under nvm and leaves neither
+# on the PATH of the shell that called it, so resolve the pinned
+# toolchain here the way bootstrap's own install step does. nvm is a
+# bash script, hence the subshell.
+run_yarn() {
+    if command -v yarn >/dev/null 2>&1; then
+        exec yarn "$@"
+    fi
+    if [ ! -s "$HOME/.nvm/nvm.sh" ]; then
+        echo "fmt-check: no yarn; run script/bootstrap first" >&2
+        exit 1
+    fi
+    exec bash -c '. "$HOME/.nvm/nvm.sh" && nvm use "$1" >/dev/null &&
+        shift && exec yarn "$@"' bash "$NODE_VERSION" "$@"
+}
+
 main() {
     cd "$ROOT"
-    yarn run prettier --check '**/*.md' --tab-width 4 --prose-wrap always
+    run_yarn run prettier --check '**/*.md' --tab-width 4 --prose-wrap always
 }
 
 main "$@"

script/lint

@@ -1,13 +1,23 @@
 #!/bin/sh
-# script/lint: run the linter.
+# script/lint: run the linter. Linting is a phase of the Dockerfile and
+# this builds that phase alone; the linter is never installed or run on
+# a developer host, where a shared result cache and a host-global lock
+# make its answer untrustworthy.
+#
+# The phase is not the last stage in the file, so it is built only when
+# --target names it. --no-cache because a cached lint layer is a lint
+# that did not run. The tag makes each build replace the previous image
+# instead of leaving a dangling one behind.
 set -eu
 
-ROOT="$(cd "$(dirname "$0")/.." && pwd -P)"
+SCRIPT_DIR="$(cd "$(dirname "$0")" && pwd -P)"
+ROOT="$(cd "$SCRIPT_DIR/.." && pwd -P)"
 
 main() {
     cd "$ROOT"
-    echo "Linting markdown files..."
+    docker build --no-cache \
-    yarn run prettier --check '**/*.md' --tab-width 4 --prose-wrap always
+        --target lint \
+        -t "$("$SCRIPT_DIR/projectname")-lint" .
 }
 
 main "$@"

script/test

@@ -1,12 +1,19 @@
 #!/bin/sh
-# script/test: run the test suite.
+# script/test: run the test suite. Testing is a phase of the Dockerfile
+# and this builds that phase alone, on the same terms as script/lint:
+# --target because a phase that is not the last stage is built only when
+# named, --no-cache because a cached test layer is a test that did not
+# run, and a tag so each build replaces the previous image.
 set -eu
 
-ROOT="$(cd "$(dirname "$0")/.." && pwd -P)"
+SCRIPT_DIR="$(cd "$(dirname "$0")" && pwd -P)"
+ROOT="$(cd "$SCRIPT_DIR/.." && pwd -P)"
 
 main() {
     cd "$ROOT"
-    echo "No tests defined."
+    docker build --no-cache \
+        --target test \
+        -t "$("$SCRIPT_DIR/projectname")-test" .
 }
 
 main "$@"