sneak/prompts
1
0
Fork 0
Code Issues 18 Pull Requests 4 Actions Packages Projects Releases Wiki Activity

Compare commits

base: sneak:417f142a9fd29c720cfb2c13aa0b8bee9b7d83a8
Branches Tags
sneak:main sneak:fix/25-gomodguard-v2 sneak:next sneak:go-test-count1 sneak:lint-policy-reconcile sneak:llm-prose-tells-v10 sneak:llm-prose-tells-merge-pass sneak:llm-prose-tells-final sneak:llm-prose-tells-checklist-pass sneak:llm-prose-tells-updates sneak:add-llm-prose-tells sneak:improve-quickstart-commands
..
compare: sneak:next
Branches Tags
sneak:fix/25-gomodguard-v2 sneak:main sneak:next sneak:go-test-count1 sneak:lint-policy-reconcile sneak:llm-prose-tells-v10 sneak:llm-prose-tells-merge-pass sneak:llm-prose-tells-final sneak:llm-prose-tells-checklist-pass sneak:llm-prose-tells-updates sneak:add-llm-prose-tells sneak:improve-quickstart-commands

6 Commits
417f142a9f ... next

Author SHA1 Message Date
sneak
cc6a5a00e7 Run every lint in a container via Dockerfile.lint (closes #40)
All checks were successful
check / check (push) Successful in 9s
Details 
script/lint runs the linter directly when it is already inside a
container and otherwise builds Dockerfile.lint, so the linter never runs
on a developer host. That closes three host-only mechanisms: the result
cache golangci-lint keys on file content rather than location, which
produced a confirmed false green and findings reported against other
checkouts; the host-global $TMPDIR/golangci-lint.lock, which fails a run
in a way no caller can distinguish from findings; and host/container
version skew, which hid thirteen findings on one repo.

Detection is on LINT_IN_CONTAINER=1, set by every Dockerfile, and on
nothing else. The two directions are not symmetric: a false negative
inside a container attempts a nested docker build, finds no daemon and
fails loudly, while a false positive on a host silently lints there,
which is the defect this issue exists to kill. /.dockerenv is therefore
rejected even as a fallback -- measured absent inside BuildKit RUN steps
and present on any host that is itself a container, so it fails in both
directions and one of them is the dangerous one.

Nothing else changes shape. The Dockerfile still runs make check,
script/check still runs test, lint and fmt-check, script/cibuild is
still a single docker build with CHECK_EPOCH and VERSION, and the Go
multistage lint stage and its COPY --from=lint ordering dependency
survive with ENV LINT_IN_CONTAINER=1 added. Dockerfile.lint is the
standalone developer-host path and carries the same CHECK_EPOCH guard,
with the ARG below the dependency layer so only the lint re-runs.

The script/bootstrap golangci-lint install and the per-checkout
GOLANGCI_LINT_CACHE/TMPDIR wrapper are deleted as superseded. Neither
has a caller left. A JS repo's yarn install stays: the rule is that no
lint verdict may come from a host invocation, not that no linter binary
may exist there, and in a repo whose formatter is its linter the
formatter necessarily runs on the host.

golangci-lint config verify is kept, on measurement. Under the pinned
v2.12.2 a bogus top-level key and a bogus key under linters.settings.lll
both pass `golangci-lint run` with exit 0 and `0 issues` while config
verify exits 3 and names them; an unknown linter name fails run and
passes config verify. It needs no network: every case reproduced
byte-identically under `docker run --network none`, in a container where
`getent hosts golangci-lint.run` exits 2.

Comment blocks were cut hard across every file this unit touches.
.dockerignore drops from 67 comment lines to 28, script/cibuild from 17
to 12, script/docker from 18 to 12, and prompts/REPO_POLICIES.md from
1182 lines to 907. What remains says why a line is load-bearing; the
discovery narratives are gone.

config verify lives in script/lint's native branch rather than in a
Dockerfile, so every path that lints inherits it: the lint stage of the
main image, which is what CI runs, as well as Dockerfile.lint. Putting
it in one Dockerfile is how the other path silently loses it.
 2026-08-10 14:05:54 +00:00
clawbot
0620416869 Give golangci-lint per-checkout cache and lock state (closes #30)
All checks were successful
check / check (push) Successful in 7s
Details 
A golangci-lint result on a host running many concurrent workers does not
reliably belong to the tree that asked for it. Two independent mechanisms,
which have repeatedly been mistaken for one:

The result cache is keyed on file content, not location, so two checkouts
of the same commit hold byte-identical files, share cache entries, and one
tree's findings are served for the other under the other tree's path. This
produced a confirmed false green as well as the loud false reds. Moving
workers from shared worktrees to their own clones does not address it —
two clones collide exactly as two worktrees did — and removes only the
foreign-path artefact that made the defect noticeable.

The concurrency lock is $TMPDIR/golangci-lint.lock (pkg/commands/run.go,
acquireFileLock), host-global and independent of GOLANGCI_LINT_CACHE, with
a five-second acquire timeout, so it fails when the host is busiest. A
private cache directory does not isolate it. Setting only the cache closes
the contamination half and leaves runs failing red on a condition that is
not a result at all.

REPO_POLICIES.md now carries the canonical Go script/lint: both variables
scoped into a .lint-cache/ directory inside the checkout, above any
container-versus-host branch so every path reaching the linter gets them;
--allow-serial-runners, which keeps the mutual-exclusion guard and queues
rather than aborting, for the same-checkout overlap TMPDIR scoping cannot
cover, with --allow-parallel-runners rejected because it deletes the
guard; and a bounded retry that treats the lock error as VOID rather than
as findings, exiting 75 on exhaustion so it is neither a pass nor a
failure. Detection is on the stderr stream and never on exit status:
findings go to stdout, so a finding quoting the lock message in source
cannot be retried away, and the exit status is not a stable discriminator
anyway. The interim void rule is recorded with the ../ clause that the
original filter missed, and with its limit stated — it catches
contamination that names foreign files, not contamination that suppresses
findings. Both checklists gained the corresponding items, since a half-fix
that sets only the cache reads as complete.

GOCACHE was measured rather than assumed and does not need isolating: with
the two variables scoped per checkout and GOCACHE shared at the host
default, each checkout reported its own paths.

Verified with the snippet extracted from the committed document and
executed as a consuming repo would adopt it, each control paired against
the pre-fix form: contamination reproduced on the pre-fix script and
absent on the adopted one; a stub linter colliding twice then clearing,
with the retry engaging and succeeding; exhaustion exiting 75 with a VOID
message; a genuine finding whose text quotes the lock message reported as
findings with no retry; and a real held lock failing the pre-fix script
with exit 3 while the adopted script, inheriting the same environment,
completed in one second.
 2026-08-09 18:51:22 +00:00
sneak
3a218497b8 Keep in-repo agent scratch out of the build context and out of git (closes #27)
All checks were successful
check / check (push) Successful in 17s
Details 
The canonical .dockerignore and .gitignore both omitted the in-repo agent
scratch directory. On this fleet that directory holds one worktree per
in-flight agent -- an entire additional checkout of the repo each -- so under
`COPY . .` all of it reached the build context and the image. Measured on this
repo before the change: five planted scratch files, at every depth beneath the
directory, all present inside a probe image built from the real context.

Three consequences, only the first of which is about size. The context inflates
by a multiple of the repo. Another session's unreviewed and sometimes
uncommitted work is copied into a build artifact. And the directory is created
and destroyed constantly by tooling, so it invalidates `COPY . .` for reasons
that have nothing to do with this repo's content -- which is the accidental
cache protection described at length in the issue thread, and the reason this
change was sequenced behind the CHECK_EPOCH bust rather than landed alongside
the rest of the .dockerignore work.

The two entries are deliberately different shapes, because the two files have
different semantics and neither is derived from the other. In .dockerignore the
entry is anchored, `.claude`, with no `**/` prefix: the directory occurs exactly
once, at the context root, and the prefixed form additionally matches any nested
directory of that name. Measured rather than argued -- the `**/`-prefixed
control was built and enumerated too, and it removes prompts/.claude/ from the
context as well, which in a repo with a legitimately named nested directory
would silently delete it from the build. In .gitignore the entry is unanchored,
`.claude/`, because a .gitignore pattern already matches at every depth;
`git check-ignore -v` confirms it covering both .claude/ and prompts/.claude/,
so a `**/` prefix there would be redundant at best, and on an anchored pattern
it would be actively wrong.

Anchoring buys that at the cost of a residual exposure, and the vendored files
now say so rather than only asserting the reason to anchor. "Occurs exactly
once, at the context root" is a property of how agents are run, not of the
tooling: the directory is created in the agent's working directory, so a
monorepo running a per-service agent in services/api/ still ships
services/api/.claude/ into the context and the image -- the exact exposure this
change exists to close, left open in the repo shape where it is likeliest. The
.dockerignore header block, the REPO_POLICIES.md bullet and both checklists
state the gap and the remedy (anchored entries for the subdirectories that have
one, or `**/.claude` once no legitimately named nested directory would be
caught). Consuming repos receive the files and not the tracker, so a caveat that
lives only in a PR body is not a caveat.

It is not case-folded the way the neighbouring secret patterns are: tooling
creates the directory in exactly one spelling, so a folded pattern would add no
coverage. The earlier justification -- that a miss costs bloat rather than
exposure -- is gone, because it contradicted this issue's own framing, in which
the cost of a miss is unreviewed work in an image layer.

The second half of this change is the consequence that ships broken silently.
Excluding .git means `git describe` cannot run in any build stage, and it fails
quietly there rather than erroring: `-X main.Version=` comes out empty, the
binary reports no version, and the build still exits 0. The Go template in
REPO_POLICIES.md had `ARG VERSION=dev` and never said where VERSION came from,
which is precisely the gap a reader fills in with `git describe` inside the
build. It now says: computed on the host, threaded in with `--build-arg
VERSION=...`, shown as a complete command rather than as two rules each
documenting half of one. script/docker and script/cibuild do it, with the same
discipline the epoch already has -- assignment on its own line, because a
failing command substitution inside an argument does not trip `set -e`, plus a
non-empty fallback so a build from an export with no .git reports `unknown`
rather than an empty string that reads as a successful version.

That fallback is applied in exactly one place, and that place can actually
execute. `git describe ... || true` leaves the value empty when it fails, and
the `[ -n "$version" ]` line is what substitutes `unknown`. Folding the fallback
into the substitution as `|| echo unknown` would have left the guard unreachable
-- harmless in itself, but a guard that cannot fire is indistinguishable from
one that works to every repo that copies it, and canonical text should not carry
a check that is decorative. Measured firing in both sh and dash: not a git
repository -> unknown, repository with no commits yet -> unknown, this
repository -> the describe output. The checklist now carries the guard line as
well, so the vendored guidance and the vendored script no longer disagree.

The scripts pass VERSION unconditionally rather than growing a per-repo variant.
This repo's Dockerfile declares no `ARG VERSION`, and BuildKit was measured
accepting the unconsumed arg silently -- no warning, no cache effect, confirmed
by the paired runs below in which the bootstrap layer still caches. The
alternative, leaving it to each repo, reintroduces the trap: a repo that needs a
version and finds no VERSION in its scripts writes `git describe` into the
Dockerfile, which is the failure being closed.

The same correction reaches the two Go documents that carry the GOLDFLAGS
pattern, since a `$(shell git describe)` evaluated inside a build stage is
exactly this empty version. Both are now `?=`, and their comments say precisely
when that matters: where a build stage compiles by invoking make, `ARG VERSION`
puts the value in the environment and `?=` defers to it, whereas the canonical
Go template compiles with `go build` directly and uses the Makefile on the host
only -- still `?=`, so that a repo which later moves its build behind make does
not silently start shipping an empty version. Leaving them as `:=`
would have left the corpus telling a reader one thing in the policy and the
opposite in the styleguide.

Both repo checklists gain the entries too. They are what an agent reads while
writing these files, so they are where the wrong shape actually gets written:
the .gitignore item is the one an existing repo never re-fetches, and it now
names `.claude/` explicitly along with the warning not to prefix it.

Verification, by enumerating a probe image rather than by reading the patterns.
Standalone minimal Dockerfile held outside the context, `--no-cache` scoped to
that one image, no prune of any kind. Before: all five planted scratch files in
the image, 42 files total. After: zero, 37 files total, with README.md,
script/check, prompts/NEW_REPO_CHECKLIST.md and a planted probe_src/app.md all
still present as positive controls, so the exclusion is a real exclusion and not
a COPY that stopped copying. Transferred context fell from 161.86kB to 68.82kB,
recorded as corroboration only: BuildKit reports a delta, not a total, and an
earlier run in this repo transferred 2.18kB while shipping 43 files.

The CHECK_EPOCH verification was re-run under the changed context, because the
context moved underneath the earlier measurement. Two consecutive script/cibuild
runs on an unchanged tree: run 1 in 17.07s, run 2 in 5.73s, both executing the
check layer with a distinct epoch and real prettier output from both lint and
fmt-check. The `RUN script/bootstrap` layer is CACHED in run 2, which is the
validity control -- it proves no concurrent prune landed between the runs and
that no --no-cache path was taken, so the check layer executing is the bust
working rather than a cold cache.

Planted files were removed afterwards and their absence confirmed against the
filesystem with `find`, not against `git status`, which cannot see them once
.gitignore covers the directory -- the same blind spot that made the earlier
secret exposure invisible.
 2026-08-09 17:13:10 +00:00
sneak
fd78aeb003 Keep secrets out of the Docker build context at every depth (closes #29)
All checks were successful
check / check (push) Successful in 21s
Details 
The canonical .dockerignore was three lines -- .git, node_modules, .DS_Store
-- while the canonical Dockerfile does `COPY . .`, so a developer's local
.env, *.pem or *.key was shipped into the build context and could land in an
image layer. Nothing surfaced it because .gitignore covers those patterns, so
the files are invisible to every git-based check.

The obvious repair, copying .gitignore's secret patterns across, is worse than
the gap it closes. .dockerignore does not use .gitignore semantics: Docker
matches with moby/patternmatcher, which is filepath.Match semantics plus a `**`
extension compiled to a regexp. `*` does not cross `/`, and a pattern without
a leading `**/` is anchored at the build-context root. A file listing .env,
*.pem and *.key therefore reads as solved, reviews as solved, and protects
only the repository root, while config/.env and certs/server.key still ship.
The three-line file at least invited scrutiny; the transplanted form
manufactures confidence and stops anyone looking.

So every depth-independent pattern here carries the `**/` prefix and only
genuinely root-anchored entries stay unprefixed. `**/node_modules` fixes a
defect the three-line file had today for any nested node_modules,
independently of the secret exposure.

Coverage is not limited to the three patterns the issue names, because the
enumeration found more shapes reaching the image. `**/*.env` covers the
prod.env / local.env convention, which the .env and .env.* spellings miss
entirely; `.envrc` is a secrets file by direnv convention; *.p12 and *.pfx are
bundles carrying private keys; and id_rsa, id_dsa, id_ecdsa and id_ed25519 are
the extensionless SSH keys that were covered before only when someone happened
to append a .key suffix.

Matching is case-sensitive, so `**/*.key` does not match certs/SERVER.KEY,
which is reachable on the case-insensitive filesystems most laptops use.
Doubling each pattern with an ALL-CAPS twin is not the fix: measured against
the planted set it still ships certs/Server.Key and certs/Ca.Pem while reading
as though case were handled, which is this issue's failure mode restated in a
new place. The matcher supports character ranges, so every secret name is
written that way -- `**/*.[kK][eE][yY]`, `**/*.[pP][eE][mM]`,
`**/.[eE][nN][vV][rR][cC]`, `**/[iI][dD]_[rR][sS][aA]` and the rest. The
extensionless SSH keys and .envrc are folded for the same reason the
extensions are: on the very filesystems that make SERVER.KEY reachable, direnv
reads .ENVRC and ssh reads ID_RSA, so exempting them would have contradicted
the rule that justifies the folding. Since `*` matches the empty string,
`**/*.[eE][nN][vV]` already covers a bare .ENV and no separate literal .env
entry is needed; the literal one is gone rather than left to imply that case
is unhandled there.

Deliberately not covered: bare `key` and `pem` filenames, which no tool
produces and which collide with legitimate paths (a `**/key` pattern would
delete an internal/key/ package directory from the context); `**/id_*`, which
would match ordinary source such as id_generator.go and ID_MAP.go; and *.crt
and *.cer, which are public certificates rather than secrets and are sometimes
a legitimate build input. Each of those is planted as a positive control and
verified present in the image after the change.

The one predictable false positive is a committed env template: `**/*.env`
excludes example.env. The remedy travels with the file rather than living in a
review thread -- the header comment and the policy both say to re-include it
with a negation, `!docs/example.env`, and never to delete the pattern, which
would reopen the exposure for everything else it covers.

The OS and editor patterns are included on their own merits rather than by
mirroring .gitignore. None of them is ever a build input, and editor state in
particular churns under a developer's hands, so each one is a source of
`COPY . .` invalidation carrying no information about the source tree. Now
that the checks are keyed on CHECK_EPOCH rather than on accidental context
churn, there is no reason left to keep churn in the context. Language build
artifacts are deliberately absent: they are per-repo, and the file's header
comment tells consuming repos to add their own host-built binaries, which is
the case that actually bites -- a host `make build` drops a multi-megabyte
artifact into the context where .gitignore hides it from every git-based
check. That comment gives the anchored form explicitly, `/myapp` rather than
`**/myapp`, because the prefixed spelling also matches cmd/myapp/ and deletes
the package directory. The header comment is the only part of this guidance a
consuming repo actually receives, since it is vendored with the file.

.gitignore is untouched. Its semantics are the inverse: an unanchored pattern
already matches at any depth, so `**/`-prefixing it produces a file that is
wrong in a way that looks careful. That asymmetry is why "derive one from the
other" was the wrong instruction, and it is now written down in
REPO_POLICIES.md in both directions, together with the case-sensitivity rule,
the negation remedy, and the requirement to verify by enumerating the image
rather than by reading the patterns. Every consuming repo inherits
.dockerignore by copy, so the trap has to live where the next person looks,
not only be fixed once here. Both repo checklists gain the same requirements.

Verified by planting 130 secret files -- twenty-six name shapes, including
every capitalisation of .env, .env.*, prod.env, .envrc, id_rsa, id_ed25519,
ca.pem, server.key, bundle.p12 and bundle.pfx, at five depths from the context
root to a/b/c -- alongside nine positive controls, then building a standalone
probe image doing `COPY . .` and listing what actually landed inside it. The
patterns as first written leaked 35 of the 130: every .ENVRC, .Envrc, ID_RSA,
Id_Rsa, ID_ED25519, .ENV.PRODUCTION and .Env.Local, at all five depths. A
lowercase-only control leaks 80, so the probe is not vacuous. After this
change: zero, with all nine controls still present, including
internal/ID_MAP.go and certs/CA.CRT.

Transferred-context size is recorded but load-bearing on nothing: an earlier
run reported 2.18kB transferred while 43 files, five of them secrets, were in
the image. BuildKit transfers only the delta from the previous build, so the
number describes the transfer and not the contents.

Planted files were removed and their absence confirmed against the filesystem
rather than against `git status`, which could not have seen them.

`make docker` re-run after the change: the check layer executed rather than
being served from cache, so the CHECK_EPOCH verification still holds under the
altered build context.
 2026-08-09 16:42:00 +00:00
clawbot
d173e69f85 Make the pinned golangci-lint actually reach the host (closes #28)
All checks were successful
check / check (push) Successful in 9s
Details 
REPO_POLICIES.md now carries the canonical script/bootstrap snippet for Go
repos alongside the .golangci.yml bullet, where the pinned linter version
already lives.

The guard it replaces, `if missing golangci-lint; then go install ...; fi`,
tests PATH presence and never version, so on any already-provisioned machine
the pin is inert and a version bump is a no-op. The Dockerfile installs
unconditionally into a clean image, so CI and local then disagree about what
the linter is: a local `make check` green while `make docker` rejects the same
commit, and a container run surfacing findings the host run cannot see.

Comparing versions alone is not enough. `go install` writes to GOBIN (or
GOPATH/bin) while callers resolve through PATH, so a shadowing binary earlier
in PATH lets the install succeed and change nothing a caller ever sees, while
bootstrap prints success. The canonical form therefore compares the installed
version against the pin, re-resolves through PATH after installing and asserts
the pin, and treats any unparseable --version output as a mismatch so the
failure direction is a redundant install rather than a skipped one.

The comparison is exact over the whole version token. A parser that stops at
the first `-` reports 2.12.2 for a host running 2.12.2-rc1, which compares
equal to a 2.12.2 pin and skips the install -- the original defect, reachable
through the comparison meant to close it, and not caught by requiring the pin
to be tagged, since a pre-release is tagged too.

When the assert fails the diagnosis is derived from the resolved path rather
than asserted: a path outside the install directory is shadowing and the
operator is told to remove it or reorder PATH; a path inside it is not, and
saying so would send them after a fault that does not exist; no resolution at
all means the install directory is simply absent from PATH. A trailing slash
on GOBIN is normalised away, since it would otherwise make the inside-the-
directory test miss and misreport shadowing.

The snippet ends with a call site, and both success paths print a confirmation
naming the version. Two definitions with no invocation are a silent no-op with
exactly the shape this change exists to close, and a success path that prints
nothing is byte-identical to that no-op: same exit status, same empty output.

The version helper ends in `|| true` so a --version that exits non-zero cannot
kill the script through `set -e` under `set -o pipefail` before the diagnostic
is printed, which the styleguide's bash form would otherwise do.

The policy text states each of those as a requirement rather than leaving them
implicit in the code, and records why the commit-pinned `go install` ref
satisfies the hash-pinning rule: a commit hash is not a mutable tag, and the
checksum database verifies the fetch, with no repo go.sum consulted, since
`go install pkg@version` ignores the go.mod in the current directory or any
parent. Whether a go.mod tool dependency should replace that is an open
decision and is linked rather than settled here. The two version strings must
be kept in sync and must match exactly what --version prints; tagged pins are
preferred because the expected string is then derivable from the ref, rather
than because the comparison cannot handle a pseudo-version.

Verification runs the controls against the block as a consuming repo would
adopt it, pasted into a script/bootstrap-shaped file and executed, rather than
sourcing it and calling the function directly.

The node and yarn handling described earlier in the document is untouched.
 2026-08-09 16:01:36 +00:00
clawbot
51c394552e Bust the Docker check-layer cache with a per-invocation CHECK_EPOCH (closes #26)
All checks were successful
check / check (push) Successful in 7s
Details 
script/cibuild was a plain `docker build .`, and the Dockerfile does
`COPY . .` followed by `RUN make check`. Docker invalidates a COPY layer
only when the copied content changes, so on an unchanged tree the check
layer was served from cache, the suite never ran, and the build still
exited 0. Measured here: run 1 took 18.5s and ran the suite; run 2 on a
byte-identical tree took 0.286s with `RUN make check` CACHED.

script/cibuild and script/docker now assign a per-invocation nonce on its
own line and pass it as --build-arg CHECK_EPOCH. The Dockerfile declares
ARG CHECK_EPOCH, guards it with `[ -n "$CHECK_EPOCH" ] || exit 1`, and
expands it into the check command. Post-fix, two consecutive runs both
execute make check (17.4s / 8.1s) with `RUN script/bootstrap` still
CACHED, so dependency layers are untouched and the build ceiling is not
at risk.

The guard is what makes a bare `docker build .` — the command
REPO_POLICIES named verbatim — fail closed rather than reuse the empty
and therefore stable cache key; verified failing in 0.455s. Holding the
epoch constant restores the false green (run 2 fully CACHED), which pins
the varying value as the operative mechanism rather than a coincidence.

Because the guard references $CHECK_EPOCH it is itself value-keyed:
BuildKit renders the epoch into that layer's description and re-runs the
layer when the value changes. Each stage therefore has two independent
invalidation points, the guard and the expansion, and the guard always
precedes the check RUN. Both are kept and the prose now records this;
the expansion remains defence in depth and is what puts the epoch in the
build log.

The false guarantee was org-canonical text in more than one document, so
it is corrected everywhere it appeared rather than only where the issue
first found it. REPO_POLICIES.md carried it in two places, and its Go
multistage template had check steps in two stages; ARG is stage-scoped,
so both stages get the treatment or the fleet inherits the half-fixed
shape. CODE_STYLEGUIDE_GO.md restated the guarantee for the bare command
this change makes fail closed. NEW_REPO_CHECKLIST.md specified the
pre-fix script/cibuild verbatim, so every new repo would have been born
with the false green, and EXISTING_REPO_CHECKLIST.md ended on a
`docker build` acceptance item that the guard makes unsatisfiable by
design — an agent working that checklist would have been led to delete
the guard to tick the last box. Both checklists' Dockerfile criteria were
also satisfiable by a Dockerfile whose check layers are still frozen, and
now require the ARG and guard in every check-running stage.

REPO_POLICIES.md's own Dockerfile criterion carried that same incomplete
form; it is tightened by cross-reference to the CHECK_EPOCH rule rather
than by duplicating the canonical block. The Go template's Key points
gain a caveat that the cache-bust turns the `COPY --from=lint` no-op into
a content-cache hit, so a repo using a file-dependency trick for stage
ordering must re-prove that ordering on a warm cache after adopting it.
That was re-proved in another repo in the org which uses the trick with a
marker file, where the ordering held; the caveat states explicitly that
it was not verified here, this repo being single-stage with no lint stage
to order against.
 2026-08-09 15:15:03 +00:00
14 changed files with 951 additions and 136 deletions
Expand all files Collapse all files

59
.dockerignore
View File

@@ -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 the 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-*

6
.gitignore vendored
View File

@@ -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/

26
Dockerfile
View File

@@ -3,26 +3,26 @@ FROM node@sha256:e4bf2a82ad0a4037d28035ae71529873c069b13eb0455466ae0bc13363826e3
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.
# Makes script/lint run the linter directly rather than building
# Dockerfile.lint, which would need a docker daemon here.
ENV LINT_IN_CONTAINER=1
# 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 . .
# CHECK_EPOCH is a per-invocation nonce supplied by script/cibuild and
# script/docker. Without it an unchanged tree serves this layer from
# CHECK_EPOCH is a per-invocation nonce from script/cibuild and
# script/docker; without it an unchanged tree serves this layer from
# cache and the build reports a green it never ran. ARG is stage-scoped,
# so it must be redeclared in every stage that runs checks. The guard
# makes a bare `docker build .` fail loudly instead of silently reusing
# the empty (and therefore stable) cache key. Expand the value into the
# command so the cache miss does not depend on BuildKit's handling of an
# unreferenced ARG. Both the guard and the check RUN reference the value,
# so both are value-keyed: there are two independent invalidation points
# here, not one. Keep both.
# so declare it in every stage that runs checks. The guard fails a bare
# `docker build .`, which would otherwise reuse the empty (and therefore
# stable) cache key. The value is also expanded into the check command,
# so the cache miss does not depend on BuildKit's handling of an
# unreferenced ARG; keep both references.
ARG CHECK_EPOCH
RUN [ -n "$CHECK_EPOCH" ] || exit 1
RUN echo "check epoch: ${CHECK_EPOCH}" && make check

27
Dockerfile.lint Normal file
View File

@@ -0,0 +1,27 @@
# Lint-only image, built by script/lint when it is not already inside a
# container. Linting is a build step, so a successful build is a clean
# lint, and nothing is bind-mounted, which matters when the daemon is
# remote.
#
# node 22-alpine, 2026-02-22
FROM node@sha256:e4bf2a82ad0a4037d28035ae71529873c069b13eb0455466ae0bc13363826e34
WORKDIR /app
# Makes script/lint run the linter directly instead of recursing into
# another docker build, which has no daemon here.
ENV LINT_IN_CONTAINER=1
COPY script/ script/
COPY package.json yarn.lock ./
RUN script/bootstrap
COPY . .
# ARG sits after the dependency layer so that layer stays cached and
# only the lint re-runs. The guard fails a bare `docker build
# -f Dockerfile.lint .`, which would otherwise reuse the empty (stable)
# cache key and report a lint it never ran.
ARG CHECK_EPOCH
RUN [ -n "$CHECK_EPOCH" ] || exit 1
RUN echo "lint epoch: ${CHECK_EPOCH}" && make lint

9
README.md
View File

@@ -117,14 +117,17 @@ alpine. We provide:
- `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/lint` — lint the markdown files with prettier
- `script/lint` — lint the markdown files with prettier. Inside a container
(`LINT_IN_CONTAINER=1`, set by both Dockerfiles) it runs prettier directly; on
a host it builds `Dockerfile.lint` so the linter still runs in a container
- `script/fmt` — format all markdown files with prettier (writes)
- `script/fmt-check` — check formatting (read-only)
- `script/check` — run all checks: `test`, `lint`, `fmt-check` (our own
extension)
- `script/docker` — build the Docker image, tagged via `script/projectname`
(byte-identical across repos)
- `script/cibuild` — cd to the repo root and
(byte-identical across repos); passes the same `CHECK_EPOCH` nonce as
`script/cibuild`
- `script/cibuild` — cd to the repo root, assign `epoch="$(date +%s%N)$$"`, then
`docker build --build-arg CHECK_EPOCH="$epoch" .` (what CI runs; the image
build runs `script/check`, and the per-invocation `CHECK_EPOCH` nonce is what
stops Docker serving that check from cache on an unchanged tree — a bare

99
TODO.md
View File

@@ -21,6 +21,105 @@ fmt-check, and commit.
# Completed Steps
- 2026-08-10: Moved every lint run into a container. `script/lint` now runs the
linter directly when `LINT_IN_CONTAINER=1` and otherwise builds
`Dockerfile.lint`, so the linter never runs on a developer host — closing the
content-keyed result cache that produced a confirmed false green, the
host-global `$TMPDIR/golangci-lint.lock`, and host/container version skew.
Detection is on that marker alone: a false negative inside a container fails
loudly on the missing daemon, while a false positive on a host would silently
restore host linting, so `/.dockerenv` is rejected outright — measured absent
inside BuildKit `RUN` steps and present on hosts that are themselves
containers. Everything else keeps its existing shape: `make check` still runs
in the image, `script/cibuild` is still one build, and the Go multistage lint
stage survives with `ENV LINT_IN_CONTAINER=1`. `Dockerfile.lint` carries the
same `CHECK_EPOCH` guard, with the `ARG` below the dependency layer so only
the lint re-runs. The `script/bootstrap` golangci-lint install and the
per-checkout cache/lock/`.lint-cache` wrapper are deleted as superseded; a JS
repo's `yarn install` stays, since the rule is about where a verdict comes
from, not about which binaries exist. `golangci-lint config verify` was kept
on measurement: a bogus config key passes `golangci-lint run` with `0 issues`
and fails `config verify`, and every case reproduced byte-identically under
`--network none`, so the schema is embedded and the line costs no network.
Comment blocks across the touched files were cut hard in the same pass.
- 2026-08-09: Made a golangci-lint result belong to the tree that asked for it.
REPO_POLICIES.md now carries the canonical Go `script/lint`, which gives the
linter per-checkout `GOLANGCI_LINT_CACHE` and per-checkout `TMPDIR`. The two
are separate defects and the second is the one that gets dropped: the result
cache is keyed on file content rather than location, so checkouts holding
identical files serve each other's findings under the other's path, while the
concurrency lock is `$TMPDIR/golangci-lint.lock` — host-global, independent of
the cache, and unaffected by isolating it. Moving workers from worktrees to
their own clones does not help either half; it only removes the foreign-path
artefact that made the defect visible. The lock error is retried rather than
surfaced, because it is not a result: it exits non-zero exactly as findings
do, and reporting it as findings sends a correct branch back for rework.
Detection is on the stderr stream and never on exit status, so a finding
quoting the lock message in source cannot be retried away, and exhaustion
exits 75 with a VOID message rather than passing or failing quietly.
`--allow-serial-runners` (which keeps the guard and queues) covers the
same-checkout overlap that `TMPDIR` scoping cannot; `--allow-parallel-runners`
is rejected outright. The stdout and stderr capture files are per invocation
rather than per checkout, because serialising the linter does not serialise
the shell's redirections: two runs in one checkout — the overlap the flag
exists to support — would otherwise truncate and read each other's output,
which is the same defect one layer above where it was fixed. Both checklists
gained the corresponding items, since a half-fix that sets only the cache
reads as complete. `GOCACHE` was measured and does not need isolating.
Verified with the snippet extracted from the committed document and executed
as a consuming repo would adopt it, against paired controls: contamination
reproduced on the pre-fix form and absent on the adopted one, retry engaged,
exhaustion loud, a genuine finding still reported, and a held host lock
failing the pre-fix script while leaving the adopted one untouched.
- 2026-08-09: Kept in-repo agent scratch out of the Docker build context and out
of version control. `.claude/` holds one worktree — an entire additional
checkout of the repo — per in-flight agent, and under `COPY . .` all of it was
reaching the image: another session's unreviewed, sometimes uncommitted work,
inflating the context by a multiple of the repo and invalidating `COPY` for
reasons unrelated to the repo's own content. The `.dockerignore` entry is
root-anchored, because the directory occurs exactly once where agents run at
the repo root and the `**/` form additionally deletes any nested directory of
that name — with the residual gap that follows from anchoring (a monorepo
running agents in subdirectories still ships `services/api/.claude/`) stated
in the canonical `.dockerignore`, the policy and the existing-repo checklist,
since consuming repos receive the files rather than the tracker; the
`.gitignore` entry is unanchored, because `.gitignore` patterns already match
at every depth, and each file is written to its own semantics rather than
derived from the other. Also closed the consequence that ships broken
silently: excluding `.git` means `git describe` cannot run in any build stage
and yields an empty version without erroring, so `script/docker` and
`script/cibuild` now compute the version on the host and pass
`--build-arg VERSION`, and `REPO_POLICIES.md` states where `VERSION` comes
from instead of leaving the reader to fill the gap with `git describe` inside
the build. The two Go documents that carry the `GOLDFLAGS` pattern were
corrected in the same pass, from `:=` to `?=`, since a `$(shell git describe)`
evaluated inside a build stage is exactly the empty version this closes.
Verified by enumerating a probe image before, after, and against the
`**/`-prefixed form, with a positive control and the `CHECK_EPOCH` cache
verification re-run under the changed build context.
- 2026-08-09: Closed the secret exposure in the canonical `.dockerignore`: a
developer's local `.env`, `*.pem` or `*.key` was reaching the Docker build
context under `COPY . .`, invisible to every git-based check because
`.gitignore` covers it. The patterns are written to `.dockerignore`'s own
`moby/patternmatcher` semantics — `**/`-prefixed so they hold at every depth,
which also fixes nested `node_modules` — rather than transplanted from
`.gitignore`, whose unprefixed form protects only the repository root while
reading as solved. Coverage extends past the `.env`/`.pem`/`.key` trio to the
`prod.env` convention, `.envrc`, PKCS#12 bundles and extensionless SSH keys,
every one of them case-folded with character ranges because matching is
case-sensitive and an ALL-CAPS twin per pattern still misses `Server.Key`.
`REPO_POLICIES.md` and both repo checklists now state that asymmetry and
require verification by enumerating the image rather than by reading the
patterns. Verified with a probe image before, against three naive forms
(unprefixed, lowercase-only, ALL-CAPS-doubled), and after.
- 2026-08-09: Made the pinned golangci-lint actually propagate: REPO_POLICIES.md
now carries the canonical `script/bootstrap` snippet for Go repos, which
installs when the installed version does not match the pin (the old
`if missing` guard tested PATH presence only, so pins were inert on any
provisioned machine and CI silently disagreed with local) and then re-resolves
the binary through `PATH` and fails loudly, naming the shadowing path, when
the install did not take effect — the failure mode the naive
compare-then-install fix leaves behind while reporting success.
- 2026-08-09: Fixed the false green in the canonical CI gate: `script/cibuild`
and `script/docker` now pass a per-invocation `CHECK_EPOCH` nonce, and the
`Dockerfile` (plus the Go multistage template in REPO_POLICIES.md, in both its

23
prompts/CODE_STYLEGUIDE_GO.md
View File

@@ -1,6 +1,6 @@
---
title: Code Styleguide — Go
last_modified: 2026-08-09
last_modified: 2026-08-10
---
1. Try to hard wrap long lines at 77 characters or less.
@@ -49,7 +49,20 @@ last_modified: 2026-08-09
```
```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 at all. The version is computed on
# the host by `script/docker` / `script/cibuild` and passed with
# `--build-arg VERSION=...`. If this repo's Dockerfile compiles by
# invoking make (`RUN make build`), `ARG VERSION` in that stage puts the
# value in the environment and `?=` defers to it. The canonical Go
# template in REPO_POLICIES.md instead runs `go build` directly with
# `-ldflags "... -X main.Version=${VERSION}"`, so there this Makefile is
# a host-only path — but it is still `?=`, because a repo that later
# moves the build behind make must not silently start shipping an empty
# version. See the git-describe rule in REPO_POLICIES.md.
VERSION ?= $(shell git describe --always --dirty)
BUILDARCH := $(shell uname -m)
GOLDFLAGS += -X main.Version=$(VERSION)
@@ -98,7 +111,11 @@ last_modified: 2026-08-09
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 always runs in a container, and `golangci-lint` 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. `script/cibuild` and `script/docker` should always make sure that

79
prompts/EXISTING_REPO_CHECKLIST.md
View File

@@ -1,6 +1,6 @@
---
title: Existing Repo Checklist
last_modified: 2026-08-09
last_modified: 2026-08-10
---
Use this checklist when beginning work in a repo that may not yet conform to our
@@ -24,9 +24,14 @@ 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) —
fetch from `https://git.eeqj.de/sneak/prompts/raw/branch/main/.gitignore`
if missing
- [ ] `.gitignore` is comprehensive (OS, editor, agent scratch, language
artifacts, secrets) — fetch from
`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: `.claude/` in
particular, unanchored, so agent worktrees cannot be committed by
accident. Do not give it a `**/` prefix — that is a `.dockerignore` form
and is wrong here.
- [ ] `.editorconfig` exists — fetch from
`https://git.eeqj.de/sneak/prompts/raw/branch/main/.editorconfig`
- [ ] `Dockerfile` and `.dockerignore` exist (fetch `.dockerignore` from
@@ -37,6 +42,47 @@ with your task.
`CHECK_EPOCH` rule in `REPO_POLICIES.md`. Without them the check layer is
served from cache on an unchanged tree and the build reports a green it
never ran.
- [ ] **Every stage that runs checks sets `ENV LINT_IN_CONTAINER=1`** — the lint
stage and the build stage both. This is the item an existing repo most
often fails after adopting the containerised lint: without it
`script/lint` tries to build `Dockerfile.lint` from inside a build step,
where there is no daemon.
- [ ] `Dockerfile.lint` exists and `script/lint` builds it when not already in a
container — see the containerised-lint rule in `REPO_POLICIES.md`. Base
image pinned by sha256 with a version/date comment, `ARG CHECK_EPOCH`
**after** the dependency layer with the guard below it.
- [ ] `.dockerignore` excludes the repo's own host-built artifacts (compiled
binaries, test binaries, coverage output), written root-anchored —
`/myapp`, never `**/myapp`, which would also match `cmd/myapp/`. An
existing repo is where such a binary is likeliest to already be sitting in
the build context, invisible to git.
- [ ] `.dockerignore` excludes `.claude`, root-anchored and with 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. Confirm by enumerating the image, not by reading
the file — `.gitignore` hides these from `git status` too.
- [ ] **Do agents in this repo run anywhere other than the repo root?** The
scratch directory is created in the agent's working directory, so the
canonical anchored entry misses `services/api/.claude/` in a monorepo with
a per-service agent — it still reaches the build context and the image. An
existing repo is where such a layout already exists, so check it here
rather than assuming the canonical entry covers you: add anchored entries
for the subdirectories that have one (`/services/api/.claude`), or
`**/.claude` once you have confirmed no legitimately named nested
directory would be caught.
- [ ] 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`. No stage calls `git describe`: `.dockerignore` excludes
`.git`, so it yields an empty version without failing the build. A
tag-derived version additionally needs `fetch-depth: 0` on the CI checkout
step, which clones shallow and fetches no tags by default.
- [ ] 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.
`.dockerignore` anchors an unprefixed pattern at the context root, so the
transplanted form leaves `config/.env` and `certs/server.key` in the build
context while reading as solved — see the `.dockerignore` rule in
`REPO_POLICIES.md`.
- [ ] 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`
@@ -63,6 +109,24 @@ 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` is the canonical detect-and-branch form, 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 container-versus-host branch, or a CI step calling the binary
directly. **Expected hits that are not the defect**: `script/fmt`, and in
a repo whose formatter is also its linter, `script/fmt-check`. Everything
else the grep finds is a real second path and goes.
- [ ] Detection is on `LINT_IN_CONTAINER` alone. Reject any `/.dockerenv` or
cgroup heuristic: absent in BuildKit `RUN` steps, present on hosts that
are themselves containers, and a false positive lints on the host.
- [ ] `script/bootstrap` installs no golangci-lint. Delete the block, its
version and ref variables, and its call site. A JS repo's `yarn install`
stays — it brings a linter along with every other dependency, which is
fine as long as no verdict is taken from it.
- [ ] The per-checkout lint state is gone: no `GOLANGCI_LINT_CACHE` or `TMPDIR`
exports, no `--allow-serial-runners`, and `.lint-cache/` removed from
`.gitignore` and `.dockerignore`.
- [ ] `make check` does not modify any files in the repo
- [ ] `make test` has a 30-second timeout
- [ ] `make test` runs real tests, not a no-op (at minimum, import/compile
@@ -109,6 +173,9 @@ with your task.
# Final
- [ ] `make check` passes
- [ ] `script/cibuild` succeeds (a bare `docker build .` fails closed by design,
on the `CHECK_EPOCH` guard)
- [ ] `make lint` runs twice on an unchanged tree with the lint layer `DONE`
both times, never `CACHED` and never sub-second
- [ ] `script/cibuild` succeeds (a bare `docker build .` or
`docker build -f Dockerfile.lint .` fails closed by design, on the
`CHECK_EPOCH` guard)
- [ ] Commit and merge fixes before starting your actual task

11
prompts/GO_HTTP_SERVER_CONVENTIONS.md
View File

@@ -1,6 +1,6 @@
---
title: Go HTTP Server Conventions
last_modified: 2026-02-22
last_modified: 2026-08-09
---
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. See the git-describe rule in REPO_POLICIES.md.
VERSION ?= $(shell git describe --tags --always)
BUILDARCH := $(shell go env GOARCH)
build:

86
prompts/NEW_REPO_CHECKLIST.md
View File

@@ -1,6 +1,6 @@
---
title: New Repo Checklist
last_modified: 2026-08-09
last_modified: 2026-08-10
---
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
@@ -52,15 +56,44 @@ 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`
- Extend `.dockerignore` with the repo's own host-built artifacts, giving
every depth-independent pattern a `**/` prefix — but write a repo-root
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. See the `.dockerignore` rule in
`REPO_POLICIES.md`. The canonical file's `.claude` entry is anchored for
the same reason as a repo-root binary; leave it that way, but note it only
covers agents running at the repo root — if this repo will run them in
subdirectories, `services/api/.claude/` is not excluded and 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=...`. `ARG VERSION=dev` is
declared in the stage that compiles, and **no stage calls `git describe`**
— `.dockerignore` excludes `.git`, so it yields an empty version without
failing the build.
- All Dockerfiles must run `make check` as a build step, and every stage
containing a check-running `RUN` must declare `ARG CHECK_EPOCH` with the
`RUN [ -n "$CHECK_EPOCH" ] || exit 1` guard immediately below it — see the
`CHECK_EPOCH` rule in `REPO_POLICIES.md`. Without them the check layer is
served from cache on an unchanged tree and the build reports a green it
never ran.
- Every stage that runs checks sets `ENV LINT_IN_CONTAINER=1`, so
`script/lint` runs the linter natively instead of trying to build
`Dockerfile.lint` where there is no daemon.
- Go repos: separate `lint` stage on the `golangci/golangci-lint` image,
with `COPY --from=lint /src/go.sum /dev/null` in the build stage to force
the ordering. Re-prove that ordering warm after adopting `CHECK_EPOCH`.
- Server: also builds and runs the application
- Non-server: brings up dev environment and runs `make check`
- Image pinned by sha256 hash with version/date comment
- [ ] `Dockerfile.lint` — the lint-only image `script/lint` builds when it is
not already inside a container. Sets `ENV LINT_IN_CONTAINER=1`; same
`ARG CHECK_EPOCH` + guard + expanded-value discipline as above, with the
`ARG` **after** the dependency layer so only the lint re-runs. Base image
pinned by sha256 with a version/date comment. Copy from
`REPO_POLICIES.md`.
- [ ] Gitea Actions workflow at `.gitea/workflows/check.yml` that runs
`script/cibuild` on push — reference
`https://git.eeqj.de/sneak/prompts/raw/branch/main/.gitea/workflows/check.yml`
@@ -87,7 +120,15 @@ are thin shims calling them. Model scripts:
then `install-precommit`, plus repo-specific init
- [ ] `script/test` / `make test` — runs real tests, not a no-op (30-second
timeout)
- [ ] `script/lint` / `make lint` — runs linter
- [ ] `script/lint` / `make lint` — runs the linter directly when
`LINT_IN_CONTAINER=1`, otherwise `epoch="$(date +%s%N)$$"` on its own line
then `docker build --build-arg CHECK_EPOCH="$epoch" -f Dockerfile.lint .`.
No lint verdict may come from a host invocation. Copy from
`REPO_POLICIES.md`. Detect on `LINT_IN_CONTAINER` only — never
`/.dockerenv`, which is absent in BuildKit `RUN` steps and present on
hosts that are themselves containers. Without the nonce this exits 0 on an
unchanged tree having linted nothing; without `-f Dockerfile.lint` it
builds the main image and lints nothing at all.
- [ ] `script/fmt` / `make fmt` — formats code (writes)
- [ ] `script/fmt-check` / `make fmt-check` — checks formatting (read-only)
- [ ] `script/check` / `make check` — runs `test`, `lint`, `fmt-check`; must not
@@ -95,14 +136,33 @@ are thin shims calling them. Model scripts:
- [ ] `script/projectname` — outputs the project name (used by `script/docker`
for the image tag)
- [ ] `script/docker` / `make docker` — builds Docker image, tagged via
`script/projectname` (byte-identical across repos); assigns
`epoch="$(date +%s%N)$$"` on its own line and passes
`--build-arg CHECK_EPOCH="$epoch"`
- [ ] `script/cibuild` — cd to repo root, assign `epoch="$(date +%s%N)$$"` on
its own line, then run `docker build --build-arg CHECK_EPOCH="$epoch" .`
(what CI runs). The build arg is mandatory: see the `CHECK_EPOCH` rule in
`REPO_POLICIES.md` for why each element is load-bearing. A bare
`docker build .` fails closed by design.
`script/projectname` (byte-identical across repos); carries the same three
version lines as `script/cibuild` below, and passes
`--build-arg CHECK_EPOCH="$epoch"` and `--build-arg VERSION="$version"`
- [ ] `script/cibuild` — cd to repo root, then, each on its own line:
```sh
epoch="$(date +%s%N)$$"
version="$(git describe --tags --always --dirty 2>/dev/null || true)"
[ -n "$version" ] || version="unknown"
docker build \
--build-arg CHECK_EPOCH="$epoch" \
--build-arg VERSION="$version" \
.
```
(what CI runs). Both build args are mandatory, and both assignments must be
on their own line: a failing command substitution inside an argument does
not trip `set -e`, so the inline form degrades silently to an empty
constant. The `[ -n "$version" ]` line is a live check that fires on an
export with no `.git` and on a repo with no commits — keep it, and do not
collapse it into `|| echo unknown`, which makes it unreachable. See the
`CHECK_EPOCH` and git-describe rules in `REPO_POLICIES.md` for why each
element is load-bearing. A bare `docker build .` fails closed by design, and
so does a bare `docker build -f Dockerfile.lint .`. The image runs
`make check`, which includes lint, so `script/cibuild` needs no separate
lint step.
- [ ] `script/precommit` — called by the pre-commit hook; runs `script/check`
- [ ] `script/install-precommit` — installs the pre-commit hook that runs
`script/precommit`
@@ -113,7 +173,11 @@ are thin shims calling them. Model scripts:
# 4. Verify
- [ ] `make check` passes
- [ ] `make lint` demonstrably runs the linter rather than returning a cached
build: run it twice on an unchanged tree and confirm the lint layer says
`DONE`, never `CACHED`, both times
- [ ] `make docker` succeeds
- [ ] `script/cibuild` succeeds and demonstrably executes
- [ ] No secrets in repo
- [ ] No mutable image/package references
- [ ] No unnecessary files in repo root

597
prompts/REPO_POLICIES.md
View File

@@ -1,6 +1,6 @@
---
title: Repository Policies
last_modified: 2026-08-09
last_modified: 2026-08-10
---
This document covers repository structure, tooling, and workflow standards. Code
@@ -60,19 +60,21 @@ 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 --build-arg CHECK_EPOCH="$epoch" .`, where
`epoch` is a per-invocation nonce (see the `CHECK_EPOCH` rule below); the
Gitea workflow calls it. Four further scripts are our own extensions to the
standard: `script/check` runs `script/test`, `script/lint`, and
`script/fmt-check`; `script/precommit` is what the git pre-commit hook runs,
and it calls `script/check`; `script/install-precommit` installs the git
pre-commit hook (the `make hooks` target shims to it); and
`script/projectname` (literally that filename) simply outputs the project's
name. Scripts that need the name call `script/projectname` — e.g.
`script/docker` assembles its image tag from it — so those scripts stay
byte-identical across all repos. Repo-type-specific pre-commit extras (e.g.
`go mod tidy` verification in Go repos) belong in `script/precommit`, not in
the hook itself. Model scripts are at
repo root and runs
`docker build --build-arg CHECK_EPOCH="$epoch" --build-arg VERSION="$version" .`,
where `epoch` is a per-invocation nonce (see the `CHECK_EPOCH` rule below) and
`version` is computed on the host because `.git` is not in the build context
(see the git-describe rule below); the Gitea workflow calls it. 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).
@@ -92,37 +94,70 @@ style conventions are in separate documents:
reading the Makefile.
- Every repo should have a `Dockerfile`. All Dockerfiles must run `make check`
as a build step so the build fails if the branch is not green. For non-server
repos, the Dockerfile should bring up a development environment and run
`make check`. For server repos, `make check` should run as an early build
stage before the final image is assembled. Dockerfiles install development
prerequisites by running `script/bootstrap` rather than duplicating installs
inline; COPY `script/` and the dependency manifests (`package.json` +
`yarn.lock`, `go.mod` + `go.sum`, etc.) before running it so the bootstrap
layer stays cached until dependencies change.
as a build step so the build fails if the branch is not green — the one
exception being `Dockerfile.lint`, which runs `make lint` alone because that
is its entire purpose — which requires `ARG CHECK_EPOCH` and its guard in
every stage containing a check-running `RUN`, per the `CHECK_EPOCH` rule
below. Without them a Dockerfile satisfies this criterion while its check
layers are served from cache, so the build cannot fail on a branch that is not
green.
**Every Dockerfile must also set `ENV LINT_IN_CONTAINER=1`**, above the
checks. `script/lint` builds `Dockerfile.lint` when it is not already in a
container; without the marker it would try that from inside a build step,
where there is no daemon. See the containerised-lint rule below.
For non-server repos, the Dockerfile should bring up a development
environment and run `make check`. For server repos, `make check` should run
as an early build stage before the final image is assembled. Dockerfiles
install development prerequisites by running `script/bootstrap` rather than
duplicating installs inline; COPY `script/` and the dependency manifests
(`package.json` + `yarn.lock`, `go.mod` + `go.sum`, etc.) before running it
so the bootstrap layer stays cached until dependencies change.
- **Every check-running `RUN` must be cache-busted with `CHECK_EPOCH`.** Docker
invalidates a `COPY` layer only when the copied content changes, so on an
unchanged tree the `RUN make check` layer is served from cache, the suite
never runs, and the build still exits 0. A sub-second `docker build` reporting
success is a cache hit, not a result. The canonical form, in **every** stage
containing a check-running `RUN`:
unchanged tree the check layer is served from cache, the suite never runs, and
the build still exits 0. A sub-second `docker build` reporting success is a
cache hit, not a result. This applies to **every** file that runs checks in a
build step — `Dockerfile` and `Dockerfile.lint` alike; a `Dockerfile.lint`
without the cache-bust is a lint that never ran, reported as a pass. The
canonical form, in **every** stage containing a check-running `RUN`, placed
**after** the dependency-install layer so that layer stays cached:
```dockerfile
ENV LINT_IN_CONTAINER=1
ARG CHECK_EPOCH
RUN [ -n "$CHECK_EPOCH" ] || exit 1
RUN echo "check epoch: ${CHECK_EPOCH}" && make check
```
`ENV LINT_IN_CONTAINER=1` belongs in every such stage too, and is the line
most often missed: without it `make check` reaches `script/lint`, which
tries to build `Dockerfile.lint` from inside a build step where there is no
daemon. See the containerised-lint rule below.
and in both `script/cibuild` and `script/docker`:
```sh
epoch="$(date +%s%N)$$"
docker build --build-arg CHECK_EPOCH="$epoch" .
version="$(git describe --tags --always --dirty 2>/dev/null || true)"
[ -n "$version" ] || version="unknown"
docker build \
--build-arg CHECK_EPOCH="$epoch" \
--build-arg VERSION="$version" \
.
```
All four elements are load-bearing; none is optional, and each guards a
failure mode that otherwise fails green:
`script/lint` needs the same nonce but is **not** this command: it builds a
different file with `-f Dockerfile.lint` and passes no version. Copy its
form from the containerised-lint rule below, not this block — a
`docker build` with no `-f` builds the main image and lints nothing.
The `VERSION` lines are there for a different reason, covered by the
git-describe rule below; they are shown here so the two rules do not each
document half a command. All four `CHECK_EPOCH` elements are load-bearing;
none is optional, and each guards a failure mode that otherwise fails green:
- `ARG` is stage-scoped, so a single declaration leaves the other check
stages frozen while the fix reviews as complete. Declare it in every stage
that runs checks, immediately above the first such `RUN`.
@@ -130,8 +165,8 @@ style conventions are in separate documents:
rather than dependent on BuildKit's handling of an unreferenced `ARG`, and
it puts the epoch in the build log. The guard is itself value-keyed, for
the same reason: it references `$CHECK_EPOCH`, so BuildKit renders the
epoch into that layer's description (observed as
`RUN [ -n "1786287053..." ] || exit 1`) and re-runs it whenever the value
epoch into that layer's description (rendered as
`RUN [ -n "<epoch>" ] || exit 1`) and re-runs it whenever the value
changes. Each stage therefore has two independent invalidation points, and
the guard always precedes the check `RUN`. Keep both: the expansion is
defence in depth, and it is what makes the epoch visible in the build
@@ -149,27 +184,189 @@ style conventions are in separate documents:
concurrent invocations would collide.
This invalidates the check layers and everything after them while leaving
`go mod download`, `script/bootstrap`, and the pinned toolchain install
cached, so it does not push against the five-minute Docker build ceiling.
Blanket `--no-cache` also works but is wasteful and can blow that ceiling.
`go mod download` and `script/bootstrap` cached, so it does not push against
the five-minute Docker build ceiling. Blanket `--no-cache` is not an
acceptable substitute: it also busts the dependency layer, so every run
reinstalls dependencies over the network instead of only the first and those
after a manifest change. Never reach for `docker builder prune` — the build
cache is shared with every other build on the host.
- **Every lint run happens in a container.** `script/lint` runs the linter
directly when it is already inside one, and otherwise builds `Dockerfile.lint`
so that it is. Either way the linter never runs on a developer host, where its
answer is not trustworthy:
- **Confirmed false green.** golangci-lint keys cached results on file
**content, not location**, so a second checkout of the same commit serves
its findings. One implementer reported `0 issues` on a branch genuinely
red with a `goconst` finding. Own-clones-instead-of-worktrees does not
help; two clones are byte-identical exactly as two worktrees were.
- **False reds**: findings reported against other checkouts and against
worktrees already deleted; in one case 399 issues returned to a clean
clone that genuinely lints 0.
- **Lock contention indistinguishable from findings.** golangci-lint flocks
`$TMPDIR/golangci-lint.lock` (`pkg/commands/run.go`, `acquireFileLock()`),
host-global and independent of `GOLANGCI_LINT_CACHE`, 5-second timeout. It
prints `parallel golangci-lint is running`, analyzes nothing, exits
non-zero. Not fixed by per-cache isolation — measured.
- **Version skew**: a host linter differing from the pinned one, with the
container surfacing thirteen findings the host missed.
A container has its own cache, its own `TMPDIR` and a binary pinned by
digest, so none of it is reachable. This supersedes the per-checkout
`GOLANGCI_LINT_CACHE`/`TMPDIR` wrapper, which existed only to make a host
run trustworthy; delete it on adoption.
The canonical `script/lint`, whose executable lines are the same in every
repo apart from the native lint command:
```sh
#!/bin/sh
# script/lint: run the linter. Inside a container, run it directly; on a
# host, build Dockerfile.lint so it runs in one anyway.
#
# LINT_IN_CONTAINER is set by this repo's Dockerfiles and is the ONLY
# accepted signal. Do not add a /.dockerenv fallback: it is absent inside
# BuildKit RUN steps and present on hosts that are themselves containers,
# so it both misses and false-positives — and a false positive silently
# restores host linting.
set -eu
ROOT="$(cd "$(dirname "$0")/.." && pwd -P)"
main() {
cd "$ROOT"
if [ "${LINT_IN_CONTAINER:-}" = "1" ]; then
# config verify lives here, not in a Dockerfile, so every path
# that lints inherits it — the lint stage of the main image as
# well as Dockerfile.lint. Duplicating it into each Dockerfile
# is how one of them silently loses it.
golangci-lint config verify --config .golangci.yml
exec golangci-lint run --config .golangci.yml ./...
fi
# Own line, and `$$` because busybox `date` drops %N silently.
# Without a fresh nonce the lint layer is cached and this exits 0
# having linted nothing.
epoch="$(date +%s%N)$$"
docker build \
--build-arg CHECK_EPOCH="$epoch" \
-f Dockerfile.lint \
.
}
main "$@"
```
and `Dockerfile.lint`, the standalone path for a developer host:
```dockerfile
# Lint-only image, built by script/lint when not already in a container.
# golangci/golangci-lint:v2.12.2 (Debian-based), 2026-08-07
FROM golangci/golangci-lint@sha256:5cceeef04e53efe1470638d4b4b4f5ceefd574955ab3941b2d9a68a8c9ad5240
WORKDIR /src
ENV LINT_IN_CONTAINER=1
COPY go.mod go.sum ./
RUN go mod download
COPY . .
# ARG after the dependency layer so only the lint re-runs.
ARG CHECK_EPOCH
RUN [ -n "$CHECK_EPOCH" ] || exit 1
RUN echo "lint epoch: ${CHECK_EPOCH}" && make lint
```
Load-bearing properties:
- **Detection rests on `LINT_IN_CONTAINER=1` and nothing else.** Every
Dockerfile in the repo sets it; a host does not. The asymmetry is the
whole design: a **false negative** inside a container tries a nested
`docker build`, finds no daemon and fails loudly, while a **false
positive** on a host silently lints there — the exact defect this rule
exists to kill. So the signal must be one only our own images can produce.
`/.dockerenv` is not such a signal and must not be used, even as a
fallback: measured, it is **absent** inside BuildKit `RUN` steps and
**present** on any host that is itself a container, which is the common
case for CI runners and agent sandboxes. It fails in both directions, and
one of them is the dangerous one.
- **`CHECK_EPOCH`, not `--no-cache`.** `docker build -f Dockerfile.lint .`
on an unchanged tree returns a sub-second cached success having linted
nothing. The `ARG` goes **after** the dependency layer so only the lint
re-runs; `--no-cache` would also reinstall dependencies on every lint.
- **Non-Go repos get the same pattern around their own linter** — `eslint`,
`ruff`, `prettier`, `shellcheck`. Only the base image and the native lint
command change.
- **Keep `golangci-lint config verify`, put it in `script/lint`, and it
costs no network.** It goes in the native branch, not in a Dockerfile, so
the lint stage of the main image inherits it along with `Dockerfile.lint`;
putting it in one Dockerfile leaves the other path unverified. The two
commands catch disjoint classes, measured under the pinned v2.12.2: a
bogus top-level key and a bogus key under `linters.settings.lll` both pass
`golangci-lint run` with **exit 0 and `0 issues`** while `config verify`
exits 3 and names them; an invalid value type fails both; an unknown
linter name fails `run` and passes `config verify`. So `run` alone
silently ignores an unknown key — the mode where a threshold reads as
configured and is not applied. It needs no network: every case reproduced
byte-identically under `docker run --network none`, in a container where
`getent hosts golangci-lint.run` exits 2. The schema is embedded in the
pinned binary. Re-run that control when bumping the pin.
- **A failed `script/lint` that names no finding is not a lint result.** On
the host path `docker build` exits 1 both for findings and for a build
that never got there (daemon down, image unpullable, disk full). BuildKit
names the failing step; read it, fix the environment, re-run. Do not
record a verdict from a run that did not lint.
**Scope: this rule is about linters, and a formatter is not one.**
`script/fmt` writes your working tree, so it can only run on the host, and
`script/fmt-check` is its read-only twin. In a repo whose formatter **is**
its linter (prettier over markdown; this repo), `script/bootstrap` therefore
installs the linter on the host as an ordinary dependency and
`script/fmt-check` runs it there. That is accepted: the version is pinned in
`package.json` and installed into the repo's own `node_modules`, so there is
no shared content-keyed cache, no host-global lock and nothing to skew
against. What is forbidden is taking a **lint verdict** from it —
`script/lint` stays the only source of one. A repo auditing itself will see
those hits and should leave them; anything else the grep finds is a real
second path to the linter and goes.
**What a consuming repo does to adopt this**, in order:
1. Add `Dockerfile.lint`.
2. Replace `script/lint` with the form above, with its own native lint
command.
3. Add `ENV LINT_IN_CONTAINER=1` to **every** stage of every Dockerfile that
runs checks — the lint stage and the build stage both.
4. Delete any golangci-lint install from `script/bootstrap`, with its
version and ref variables and its call site. No lint verdict comes from
the host any more, so it can only reintroduce version skew. A JS repo's
`yarn install` stays.
5. Delete the per-checkout lint state: `GOLANGCI_LINT_CACHE` and `TMPDIR`
exports, `--allow-serial-runners`, the retry/VOID wrapper, and
`.lint-cache/` from both `.gitignore` and `.dockerignore`.
6. Verify by running `make lint` twice on an unchanged tree: the lint layer
must be `DONE` both times, never `CACHED`. Then plant a violation,
confirm it fails naming the finding, revert. A bare
`docker build -f Dockerfile.lint .` must fail on the guard.
`script/check`, `script/cibuild`, `script/docker` and the `Dockerfile` are
unchanged by this: `make check` still runs inside the image, and
`script/lint` there takes the native path.
- **Dockerfiles must use a separate lint stage for fail-fast feedback.** Go
repos use a multistage build where linting runs in an independent stage based
on the `golangci/golangci-lint` image (pinned by hash). This stage runs
`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:
on the `golangci/golangci-lint` image (pinned by hash), so lint failures
surface in seconds rather than after a full compile. The build stage declares
an explicit dependency on it via `COPY --from=lint /src/go.sum /dev/null`,
which forces BuildKit — which runs stages in parallel by default — to finish
linting first. The canonical Go repo `Dockerfile`:
```dockerfile
# Lint stage — fast feedback on formatting and lint issues
# golangci/golangci-lint:v2.x.x, YYYY-MM-DD
FROM golangci/golangci-lint@sha256:... AS lint
WORKDIR /src
ENV LINT_IN_CONTAINER=1
COPY go.mod go.sum ./
RUN go mod download
COPY . .
@@ -182,6 +379,7 @@ style conventions are in separate documents:
# golang:1.x-alpine, YYYY-MM-DD
FROM golang@sha256:... AS builder
WORKDIR /src
ENV LINT_IN_CONTAINER=1
# Force BuildKit to run the lint stage before proceeding
COPY --from=lint /src/go.sum /dev/null
@@ -193,6 +391,9 @@ style conventions are in separate documents:
RUN [ -n "$CHECK_EPOCH" ] || exit 1
RUN echo "check epoch: ${CHECK_EPOCH}" && make test
# VERSION comes from the host via --build-arg; see the git-describe rule
# below. Never run `git describe` here: .dockerignore excludes .git, so
# it yields an empty version without failing the build.
ARG VERSION=dev
RUN CGO_ENABLED=0 go build -trimpath \
-ldflags="-s -w -X main.Version=${VERSION}" \
@@ -205,44 +406,46 @@ style conventions are in separate documents:
```
Key points:
- The lint stage uses the `golangci/golangci-lint` image directly (it
includes both Go and the linter), so there is no need to install the
linter separately.
- `COPY --from=lint /src/go.sum /dev/null` is a no-op file copy that creates
a stage dependency. BuildKit runs stages in parallel by default; without
this line, the build stage would not wait for lint to finish and a lint
failure might not fail the overall build.
- 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
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`.
- The build stage runs `make test` after compilation setup. Tests run in the
build stage, not the lint stage, because they may require compiled
- The lint stage uses the `golangci/golangci-lint` image directly (it has
both Go and the linter), so nothing needs installing. `make lint` there
runs `script/lint`, which sees `LINT_IN_CONTAINER=1` and invokes
`golangci-lint` natively instead of building `Dockerfile.lint`. Without
that `ENV` the stage would attempt a nested build and fail.
- `COPY --from=lint /src/go.sum /dev/null` is a no-op copy that exists only
to create the stage dependency; without it a lint failure might not fail
the overall build.
- **Re-prove that ordering on a warm cache after adopting `CHECK_EPOCH`.**
The cache-bust turns the no-op `COPY` into a content-cache hit, so an
ordering guarantee established cold does not automatically carry over. It
was re-proved in another org repo using the same trick and held, but that
result does not transfer by assumption — re-check it warm.
- If the project uses `//go:embed` referencing build artifacts, the lint
stage must create placeholders so the directives resolve:
`RUN mkdir -p web/dist && touch web/dist/index.html`.
- If linting needs CGO or system libraries (e.g. `vips-dev`), `apk add` them
in the lint stage.
- Tests run in the build stage, not the lint stage: they may need compiled
artifacts or heavier dependencies.
- `ARG CHECK_EPOCH` appears in **both** stages, because `ARG` is
stage-scoped: declaring it only in the lint stage leaves `make test`
frozen at the last cached result. In each stage the guard sits immediately
below the `ARG` so a bare `docker build .` fails instead of reusing the
empty cache key, and the value is expanded into the first check `RUN` so
the cache miss does not rely on BuildKit's unreferenced-`ARG` handling.
Both of those lines reference `$CHECK_EPOCH`, so both are value-keyed:
each stage is invalidated at two independent points. The later `RUN`s in
the same stage need no expansion of their own: they are already
invalidated by their busted parent layer.
frozen at its last cached result. In each stage the guard sits immediately
below the `ARG` and the value is expanded into the first check `RUN`.
Later `RUN`s in the same stage need no expansion; their parent layer is
already busted.
- `ARG VERSION=dev` is declared in the build stage and supplied by
`script/docker` and `script/cibuild`. **No stage may call
`git describe`**: `.dockerignore` excludes `.git`, so it yields an empty
version without failing. See the git-describe rule further down.
- Every repo should have a Gitea Actions workflow (`.gitea/workflows/`) that
runs `script/cibuild` (which runs
`docker build --build-arg CHECK_EPOCH="$epoch" .`) on push. The Dockerfile
runs `make check`, so a successful build implies all checks pass — but that
implication holds **only** because of the `CHECK_EPOCH` cache-bust described
above. Without it, an unchanged tree serves the check layer from cache and the
build reports a green it never earned. A bare `docker build .` fails closed by
design, on the `[ -n "$CHECK_EPOCH" ]` guard; always go through
`script/cibuild` or `script/docker`. Never accept a `script/cibuild` pass as
`docker build --build-arg CHECK_EPOCH="$epoch" --build-arg VERSION="$version" .`)
on push. The Dockerfile runs `make check`, so a successful build implies all
checks pass — but that implication holds **only** because of the `CHECK_EPOCH`
cache-bust described above. Without it, an unchanged tree serves the check
layer from cache and the build reports a green it never earned. A bare
`docker build .` fails closed by design, on the `[ -n "$CHECK_EPOCH" ]` guard;
always go through `script/cibuild` or `script/docker`. Never accept a pass as
evidence without confirming it ran: a sub-second wall time, or `CACHED` on the
check layer, means nothing was executed.
@@ -314,10 +517,140 @@ 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/`.
Fetch the standard `.gitignore` from
editor files (`.swp`, `*~`), in-repo agent scratch directories (`.claude/`,
which holds one worktree — an entire additional checkout of the repo — per
in-flight agent), language build artifacts, and `node_modules/`. Fetch the
standard `.gitignore` from
`https://git.eeqj.de/sneak/prompts/raw/branch/main/.gitignore` when setting up
a new repo.
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 — see the
next rule.
- **`.dockerignore` does not use `.gitignore` semantics, and copying patterns
across unmodified leaves secrets in the build context.** Docker matches with
`moby/patternmatcher`: Go `filepath.Match` semantics plus a `**` extension,
compiled to a regexp — plain `filepath.Match` has no `**` at all. 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; `config/.env` and
`certs/server.key` still reach the context and can land in an image layer.
That file is more dangerous than a short one with no secret patterns at all,
because it reads as solved and stops anyone looking. Give every
depth-independent pattern the `**/` prefix — `**/node_modules`,
`**/.DS_Store`, and the secret patterns in the canonical file, which are
additionally case-folded per the rule below — and leave only genuinely
root-anchored entries unprefixed: `.git`, the in-repo agent scratch directory
`.claude`, and the repo's own host-built binary. The inverse move is equally
wrong: never apply `**/` to `.gitignore`, where it is redundant and produces a
file that is wrong in a way that looks careful. Each file is written to its
own semantics; neither is derived from the other. Fetch the standard
`.dockerignore` from
`https://git.eeqj.de/sneak/prompts/raw/branch/main/.dockerignore` and extend
it with the repo's own host-built artifacts — a host `make build` that leaves
a compiled binary in the repo root puts that binary in the build context,
where `.gitignore` hides it from every git-based check. Write that binary
anchored, `/myapp` and never `**/myapp`: the prefixed form also matches
`cmd/myapp/` and deletes the package directory from the context.
- **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 with `COPY . .` the build context
inflates by a multiple of the repo, and another session's unreviewed,
sometimes uncommitted work can be copied into an image layer. The directory is
also created and destroyed constantly, so it invalidates `COPY . .` for
reasons that have nothing to do with the repo's own content. In `.gitignore`
the entry is `.claude/`, unanchored, which already matches at every depth. In
`.dockerignore` it is `.claude`, anchored and with **no** `**/` prefix: the
directory occurs exactly once **where agents run at the repo root**, and the
prefixed form would also match any nested directory of that name and delete it
from the build. It is not case-folded the way the secret patterns are, because
tooling creates it in exactly one spelling, so a folded pattern would add no
coverage.
**Known gap that comes with the anchored form.** The directory is created in
the agent's working directory, so the "exactly once, at the root" premise is
a property of how agents are run and not of the tooling. Where agents run in
subdirectories — a monorepo with a per-service agent is the ordinary case —
`services/api/.claude/` is **not** excluded by the canonical entry and still
reaches the build context and the image, which is the exposure the entry
exists to close. A repo in that shape adds its own anchored entries
(`/services/api/.claude`), or `**/.claude` once it has confirmed no
legitimately named nested directory would be caught. This is stated in the
canonical `.dockerignore` itself, since that file is what consuming repos
receive.
- **`.dockerignore` matching is case-sensitive, so cover capitalisation with
character classes rather than by doubling patterns.** `**/*.key` does not
match `certs/SERVER.KEY`, which is reachable on the case-insensitive
filesystems most laptops use. Adding an ALL-CAPS twin for each pattern is not
the fix: it still misses `Server.Key` and `Ca.Pem` while reading as though
case were handled — the same manufactured confidence as the root-anchored
form. The matcher supports character ranges, so one line covers every
spelling: `**/*.[kK][eE][yY]`, `**/*.[pP][eE][mM]`. Apply this to every secret
name, not only to extensions: the extensionless SSH keys and `.envrc` need it
for the same reason, since on the very filesystems that make `SERVER.KEY`
reachable, direnv reads `.ENVRC` and ssh reads `ID_RSA`. Note that `*` matches
the empty string, so `**/*.[eE][nN][vV]` already covers a bare `.ENV` and no
separate literal `.env` entry is needed.
- **A pattern that also catches something the build needs is re-included with a
negation, not deleted.** The canonical `**/*.[eE][nN][vV]` excludes a
committed env template such as `example.env`; a repo whose build genuinely
reads one adds `!docs/example.env` after the pattern. Deleting the pattern
instead reopens the exposure for every other file it covers.
- **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`). Reading the patterns and
agreeing they look right is exactly what lets the root-only form through. 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, so the
reported size describes the transfer and not the contents of the image.
- **Excluding `.git` means `git describe` cannot run inside any build stage, and
it fails quietly there.** The `GOLDFLAGS` version-embedding pattern assumes
`.git` is present; in a build stage there is no repository, so `git describe`
writes nothing to stdout and the `-X main.Version=` value comes out **empty**
rather than erroring. The binary then 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
# Assign on its own line: a failing command substitution inside an
# argument does not trip `set -e`, so the inline form degrades to an
# empty constant — the same silent-empty failure this rule is about.
version="$(git describe --tags --always --dirty 2>/dev/null || true)"
[ -n "$version" ] || version="unknown"
docker build \
--build-arg CHECK_EPOCH="$epoch" \
--build-arg VERSION="$version" \
.
```
`--always` makes an untagged repo yield the abbreviated commit hash instead
of failing. `|| true` keeps a failing `git describe` from tripping `set -e`
and leaves the value empty, so the `[ -n "$version" ]` line is the single
place the fallback is applied — and it is a **live** check, not defence in
depth: it fires on a build from an export with no `.git`, and on a
repository with no commits yet. Do not fold the fallback into the
substitution as `|| echo unknown`; that makes the guard unreachable, and a
guard that cannot fire is indistinguishable from one that works to everyone
who copies it. The result is non-empty by construction either way, which is
the point: an empty version reads as a successful one, while `unknown` is
visibly wrong. The Dockerfile's side is `ARG VERSION=dev` in the stage that
compiles, declared there and not inherited, because `ARG` is stage-scoped
exactly as `CHECK_EPOCH` is. Passing `VERSION` to a repo whose Dockerfile
declares no such `ARG` is silently ignored by BuildKit and costs nothing,
which is why the scripts stay byte-identical rather than growing a per-repo
variant.
One consequence for CI: the standard checkout action clones shallow and
fetches no tags, so `git describe --tags` there falls back to a bare commit
hash. A repo that embeds a tag-derived version must set `fetch-depth: 0` on
its checkout step; a repo that does not embed a version needs no change.
- **No build artifacts in version control.** Code-derived data (compiled
bundles, minified output, generated assets) must never be committed to the
@@ -336,9 +669,111 @@ style conventions are in separate documents:
- `.golangci.yml` is standardized and must _NEVER_ be modified by an agent, only
manually by the user. Fetch from
`https://git.eeqj.de/sneak/prompts/raw/branch/main/.golangci.yml`. The
canonical golangci-lint version is v2.12.2 (released 2026-05-06), installed
commit-pinned via
`go install github.com/golangci/golangci-lint/v2/cmd/golangci-lint@c0d3ddc9cf3faa61a4e378e879ece580256d76e5`.
canonical golangci-lint version is v2.12.2 (released 2026-05-06), pinned as
the image digest in `Dockerfile.lint`
(`golangci/golangci-lint@sha256:5cceeef04e53efe1470638d4b4b4f5ceefd574955ab3941b2d9a68a8c9ad5240`,
which reports
`golangci-lint has version 2.12.2 built with go1.26.2 from c0d3ddc9`). That
digest is the only pin: golangci-lint is not installed on the host by any
repo. Bumping the version means changing that one digest.
- **`script/bootstrap` must not install golangci-lint.** This supersedes the
pinned host install that used to be canonical here. `script/lint` never runs
it on the host — it either builds `Dockerfile.lint` or is already in a
container that ships the binary — so a host install has no caller, and its
only remaining effect is to put a second, independently-versioned linter where
somebody eventually runs it by hand and believes the result. Delete the block,
its version and ref variables, and its call site.
This is not a ban on host dependency installs generally. A JS or docs repo's
`script/bootstrap` runs `yarn install`, which brings its linter along with
every other dependency; that is unavoidable and fine. The rule is about a
**dedicated** linter install, and about where a verdict may come from.
**The version-enforcement principle it established still applies to any
other tool a repo pins and installs on the host**, and it is the part worth
keeping, because each of its four properties guards a failure that otherwise
reports success:
- **Compare the installed version against the pin, never test presence.** A
`if missing <tool>; then install; fi` guard tests `PATH` presence and
never version, so on any already-provisioned machine the pin is inert and
a version bump is a silent no-op. Compare the **whole** version token,
exactly: a parser that stops at the first `-` reports `2.12.2` for a host
running `2.12.2-rc1` and skips the install — the original defect,
reintroduced through the comparison meant to fix it.
- **After installing, re-resolve the binary the way callers resolve it** —
through `PATH`, not the directory the installer wrote to — and assert the
reported version is the pin. An installer that writes to `GOBIN` while a
different binary shadows it earlier in `PATH` genuinely succeeds and
changes nothing any caller sees, which is worse than no fix: it converts a
known-stale tool into one everyone believes is pinned. Run `hash -r` first
so the shell does not answer from its own lookup cache, and when the
assertion fails, name the path `command -v` found, the version it reports,
and the directory the install wrote to. Diagnose from the resolved path
rather than asserting a cause: only a path **outside** the install
directory is shadowing.
- **A mis-parse must fall through to reinstall, never to a false match.**
Absent binary, non-zero exit, empty output and unrecognised output should
all yield an empty string, which compares unequal to the pin. The failure
direction is always a redundant install, never a skipped one.
- **Call it, and say so on success.** A function defined and never called is
a silent no-op indistinguishable from success: exit 0, nothing installed,
no output. Both success branches must print a line naming the version.
Verifying such logic requires a negative control in an environment where a
shadowing binary exists earlier in `PATH` than the install target — without
it the control passes against the naive compare-then-install form too and
proves nothing — plus a mis-parse control that feeds unparseable `--version`
output and confirms a reinstall. Run those controls against the block as a
consuming repo would adopt it: pasted into a `script/bootstrap`-shaped file
that is then executed, never by sourcing it and invoking the function
yourself. Driving the function directly tests something the artifact does
not do, and it is exactly how a missing call site passes every control while
the adopted snippet does nothing.
Keep it POSIX sh: no bashisms, no arrays, no `[[`, no `grep -P`.
- **Superseded: the per-checkout `GOLANGCI_LINT_CACHE`/`TMPDIR` wrapper for
`script/lint`.** It existed only to make a host lint run trustworthy, and the
containerised-lint rule above removes the host run. Delete the wrapper, the
`--allow-serial-runners` flag, and `.lint-cache/` from both `.gitignore` and
`.dockerignore`. Two of its conclusions outlive it: **`GOCACHE` does not need
isolating** (measured — content-addressed, no foreign paths in its entries, no
global lock), and **verifying lint plumbing requires paired controls** run
against the artifact as a consuming repo would adopt it, since a control that
passes against the broken form proves nothing.
- **Interim rule for reading a lint result produced on the host, in a repo that
has not yet adopted the containerised lint above.** A lint run is **VOID**
unless both hold:
- the output contains no `parallel golangci-lint is running`, and
- no reported file path begins with `../`, and none is an absolute path
outside the tree the run was launched from.
Do not record a verdict from a void run, and do not "fix" findings in files
the change does not touch — chasing phantom findings across untouched files
puts unrelated edits into a reviewed diff, which is more expensive than the
wasted rework.
The `../` clause is the one that actually bites, and it is why a filter
keyed on `/tmp` or on absolute prefixes is not enough: golangci-lint reports
paths relative to its own resolved root rather than yours, and three of the
org's reported sightings had relative paths and would have passed such a
filter. Both clauses are needed and neither alone is sufficient — one
reproduction exited non-zero with the lock error and no foreign paths at
all, and another reported 34 well-formed findings, every one of them against
another checkout.
**State the limit of these tests rather than treating them as a guarantee.**
They catch contamination that **names** foreign files. They cannot catch
contamination that **suppresses** findings through a poisoned entry for
colliding content, which has no wall-clock tell either — **no evidence of
that mode has been observed, and nobody should go chasing it**; the point is
the reach of the tests, not a claim that the mode exists. They are a filter
for the loud mode, not a proof of soundness — which is the whole argument
for containerising the linter instead of documenting a discipline that
depends on every agent remembering to apply it. Adopt the rule above and
this one stops applying to the repo entirely.
- When pinning images or packages by hash, add a comment above the reference
with the version and date (YYYY-MM-DD).

22
script/cibuild
View File

@@ -1,20 +1,28 @@
#!/bin/sh
# script/cibuild: run the CI build. The Dockerfile runs script/check, but
# that only proves anything because CHECK_EPOCH is a fresh nonce on every
# invocation: without it Docker serves the check layer from cache on an
# unchanged tree and the build exits 0 without running the suite.
# invocation: without it Docker serves the check layer from cache and the
# build exits 0 without running the suite.
set -eu
ROOT="$(cd "$(dirname "$0")/.." && pwd -P)"
main() {
cd "$ROOT"
# Assign on its own line: a failing command substitution inside an
# argument does not trip `set -e`, which would silently degrade the
# nonce to an empty constant. `$$` is required because busybox `date`
# drops %N without erroring.
# Both assignments on their own line: a failing command substitution
# inside an argument does not trip `set -e`, so the inline form
# degrades silently to an empty constant. `$$` because busybox `date`
# drops %N without erroring. VERSION is computed here because
# .dockerignore excludes .git, so `git describe` in a build stage
# yields an empty version without failing; the guard below is the
# single place the fallback is applied.
epoch="$(date +%s%N)$$"
docker build --build-arg CHECK_EPOCH="$epoch" .
version="$(git describe --tags --always --dirty 2>/dev/null || true)"
[ -n "$version" ] || version="unknown"
docker build \
--build-arg CHECK_EPOCH="$epoch" \
--build-arg VERSION="$version" \
.
}
main "$@"

16
script/docker
View File

@@ -11,12 +11,18 @@ ROOT="$(cd "$SCRIPT_DIR/.." && pwd -P)"
main() {
cd "$ROOT"
# Assign on its own line: a failing command substitution inside an
# argument does not trip `set -e`, which would silently degrade the
# nonce to an empty constant. `$$` is required because busybox `date`
# drops %N without erroring.
# Both assignments on their own line: a failing command substitution
# inside an argument does not trip `set -e`, so the inline form
# degrades silently to an empty constant. `$$` because busybox `date`
# drops %N without erroring. VERSION is computed here because
# .dockerignore excludes .git, so `git describe` in a build stage
# yields an empty version without failing.
epoch="$(date +%s%N)$$"
docker build --build-arg CHECK_EPOCH="$epoch" \
version="$(git describe --tags --always --dirty 2>/dev/null || true)"
[ -n "$version" ] || version="unknown"
docker build \
--build-arg CHECK_EPOCH="$epoch" \
--build-arg VERSION="$version" \
-t "$("$SCRIPT_DIR/projectname")" .
}

27
script/lint
View File

@@ -1,13 +1,34 @@
#!/bin/sh
# script/lint: run the linter.
# script/lint: run the linter. Inside a container, run it directly;
# on a host, build Dockerfile.lint so it runs in one anyway. The linter
# is never run on a developer host, where a shared result cache, a
# host-global lock and a stale toolchain make its answer untrustworthy.
#
# LINT_IN_CONTAINER is set by this repo's Dockerfiles and is the ONLY
# accepted signal. Do not add a /.dockerenv fallback: it is absent
# inside BuildKit RUN steps and present on hosts that are themselves
# containers, so it both misses and false-positives — and a false
# positive silently restores host linting.
set -eu
ROOT="$(cd "$(dirname "$0")/.." && pwd -P)"
main() {
cd "$ROOT"
echo "Linting markdown files..."
yarn run prettier --check '**/*.md' --tab-width 4 --prose-wrap always
if [ "${LINT_IN_CONTAINER:-}" = "1" ]; then
exec yarn run prettier --check '**/*.md' \
--tab-width 4 --prose-wrap always
fi
# Own line, and `$$` because busybox `date` drops %N silently.
# Without a fresh nonce the lint layer is cached and this exits 0
# having linted nothing.
epoch="$(date +%s%N)$$"
docker build \
--build-arg CHECK_EPOCH="$epoch" \
-f Dockerfile.lint \
.
}
main "$@"