The E2E suite has never run in CI. repoRoot() walked up looking for
PLAN.md, but .gitignore excludes PLAN.md on purpose ("Internal planning
docs - never published with the code"), so it is absent from every fresh
checkout and the helper always failed. All twelve E2E tests died on it,
in this run and identically in #18 before any of these changes.
The marker is now the migrations directory, which is committed and is
what every caller of repoRoot is ultimately looking for.
Reproduced by exporting the tree with git archive - which yields exactly
what actions/checkout puts on the runner - and running the suite there:
it fails with the old helper and passes with the new one.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
CI gains a security job (Trivy fs: vuln, secret and misconfig) and an
image job that builds each of the four services and scans the result. The
secret scanner earns its place here: a DKIM key or Cloudflare token
committed by accident is a full compromise of every domain served.
release.yml publishes on a merge into main and on v* tags. Four services
on two architectures is eight build jobs pushing by digest, then four
publish jobs joining each pair into one manifest list, tagged latest, sha
and semver. github.repository is lowercased first because GHCR rejects an
uppercase path and this repository is lucasrguerra/LambdaMail.
Both scanners failed on first run, so this commit also fixes what they
found rather than shipping a workflow that is red on arrival:
- All four images ran as root (DS-0002), which section 17 risk R5
explicitly argues against. They now run unprivileged; protocols keeps
its eight ports below 1024 through cap_net_bind_service on the binary
rather than through being root.
- A HIGH advisory in postcss, reached only through Next. "npm audit fix
--force" would have resolved it by downgrading Next from 15 to 9, so
it is pinned with an override instead.
- The Node images copied the build stage's node_modules wholesale,
shipping typescript, eslint and vitest into production, and the base
image's own bundled npm accounted for the rest of the CVEs. Runtime
dependencies now come from npm ci --omit=dev and npm is removed from
the final layer.
- compose-up could never have passed: the proxy network is external
because Coolify owns it on the deploy host, and nothing creates it in
CI, so compose refused to start the stack at all.
Verified locally: actionlint clean, all ten action refs confirmed to
exist, Trivy fs and all four image scans exit 0, the full stack reaches
Healthy on every container, and the containers run as uid 10001/1000 with
all eight privileged ports bound and SMTP answering on 25. The arm64
build itself is unverified - this machine has no buildx - and will first
be exercised by release.yml.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
"docker compose up" could not work, which means the F0 acceptance
criterion was not being met:
- ghcr.io/rspamd/rspamd:3.9 returns "denied" and clamav/clamav:1.3 does
not exist. Both were verified against the registries; the working
references are rspamd/rspamd:3.9 and clamav/clamav:1.4.
- Rspamd received RSPAMD_REDIS_SERVERS and RSPAMD_REDIS_PASSWORD as
environment variables it never reads. Confirmed in the container that
it does not expand ${VAR} in UCL: the password reached Redis as that
literal string. Greylisting and ratelimit keep their state in Redis,
so both were silently doing nothing while appearing configured. The
credential is now rendered at startup into override.d, a tmpfs, so it
touches neither the image nor a volume.
- The ratelimit module ships enabled with no rates defined, which
accepts everything. ratelimit.conf defines the per-IP, per-user and
per-recipient buckets of section 5.2.
Verified by bringing redis and rspamd up: the real password resolves, the
rates load, and a probe message returns "Action: greylist" with the state
written to Redis.
Services also gained an image: alongside build:, so the deploy host pulls
instead of compiling while "docker compose build" still works locally.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Every fix here is a case where the code and the plan disagreed, and the
code was wrong.
Certificates (section 8.2, 8.4). The parser documented that it skipped
expired entries but never checked, so an expired certificate would be
served until every handshake failed. LastReload() only advanced when
acme.json changed - and it does not change for sixty days - which made
the timestamp risk R2 monitors indistinguishable from a watcher gone
blind. Expiry is now checked on every pass, LastReload records the check
and LastChange the swap, and the <7d/<24h alerts of section 8.4 exist.
ARC (section 5). buildArcSealer resolved the DKIM key at startup, before
runDnsReconciler had provisioned one. On a first boot ARC was disabled
until someone restarted the process, which the log even admitted while
offering no way for it to start. A LazySealer binds on first use instead.
Health (section 8.4). There was no degraded state at all. /health now
reports one - still 200, because restarting cannot fix a certificate
Traefik never issued.
Bounces. DSNs were sent From postmaster@lambdamail.local, a domain that
does not resolve, so the notification was itself undeliverable. They now
carry the configured mail host, a Message-ID, and Auto-Submitted so a
vacation responder cannot loop against them.
Fan-out atomicity (section 6.1). Persist ran one transaction per
recipient, so a failure halfway through an alias fan-out left earlier
recipients with a copy while the sender was told to retry - and the retry
delivered to them twice. PersistAll commits the whole delivery or none of
it. Verified against Postgres by reverting the repository to the old
behaviour and confirming the new test fails on it.
SMTP replies. The use case encoded replies in error text ("550 5.7.1 ...")
which the adapter re-parsed by prefix, so rewording a message could turn a
permanent refusal into a retry. SmtpRejection carries the decision and the
adapter renders it.
Repository tests. These only run with TEST_DATABASE_URL, which CI does
set - so the job was red: copy and expunge seeded a 68-character digest
into a CHAR(64) column, and three others used fixed digests that collide
with the UNIQUE index on any re-run. A shared testBlobDigest() helper
fixes both; the suite now passes twice in a row against one database.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Adds the second F2 sub-project on top of the merged IMAP core: full
SEARCH criteria matching (flags, dates, size, headers, body/text,
NOT/OR), EXPUNGE (soft-delete via expunged_at), COPY, and MOVE (RFC
6851, advertised via CapMove/CapUIDPlus). Verified end-to-end with a
real go-imap/v2 client against embedded Postgres.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Merges the first F2 sub-project: a real :143 IMAP listener with Argon2id
auth, folder listing/selection, message FETCH (real ENVELOPE/BODY[] from
stored MIME), and flag STORE. Verified end-to-end with a real go-imap/v2
client against a message seeded through F1's SMTP path. See commit
9fa1929 for full details.
Adds a real :143 IMAP listener to the Go protocols service: a mailbox
owner authenticates over IMAP (Argon2id against mailboxes.password_hash,
enumeration-resistant per PLAN.md section 14.6), lists folders, selects
INBOX, fetches messages F1's SMTP path already persisted - including real
ENVELOPE/BODY[] built from the actual stored MIME bytes via go-imap/v2's
own ExtractEnvelope/ExtractBodySection helpers, not reconstructed from
denormalized columns - and sets/reads flags via STORE.
- ImapSessionUseCase orchestrates auth, folder listing/selection, message
querying, and flag mutation against new Postgres repositories (auth,
folders, message query, flags) and a local-disk blob reader.
- A github.com/emersion/go-imap/v2 (pinned v2.0.0-beta.8) Backend/Session
implements 7 real methods (Login, List, Select, Unselect, Status, Fetch,
Store) plus a correct no-op Poll; the other 10 of the interface's 18
methods (Create/Delete/Rename/Subscribe/Unsubscribe/Append/Idle/Expunge/
Search/Copy) are explicit stubs returning a clear "not yet implemented"
IMAP error - SEARCH/EXPUNGE/COPY/MOVE, IDLE/CONDSTORE, mailbox
management, and APPEND are separate follow-up sub-projects per the
design doc's agreed ordering with the user.
Verified end-to-end with a real go-imap/v2 imapclient (embedded Postgres,
no Docker needed) driving Login -> Select -> Fetch -> Store -> List
against a message seeded through F1's real SMTP use case, proving true
SMTP-write/IMAP-read interop, not just isolated unit tests.
Built via 8 sequential tasks under subagent-driven development, each
independently reviewed, plus a final whole-branch review. Two real bugs
were found and fixed before merge, both exactly the kind unit tests with
fakes cannot catch:
- Store's UID-set resolution silently no-op'd on dynamic ranges like
"UID STORE 1:*" (the common "mark all as read" operation) after an
earlier draft bypassed the folder's real message list; restored to
resolve against real messages via the same pattern Fetch already used
correctly, with tests proving both dynamic ranges and non-UID STORE.
- go-imap/v2's server calls Session.Poll() implicitly before the status
response of nearly every authenticated command (FETCH/STORE/SEARCH/
APPEND/COPY), not just IDLE as assumed - a hard-stubbed Poll broke every
real command through this path; fixed to a correct no-op.
- List had zero test coverage despite being committed scope; added unit
tests for wildcard/delimiter matching plus a real E2E LIST assertion.
Known follow-up, not blocking: a pre-existing test-fixture hygiene gap
(hardcoded content_sha256 blob hashes with no cleanup, shared across
several Postgres integration test files including one new to this
branch) causes local-dev test collisions on repeated runs against a
persistent database - invisible to CI (fresh Postgres per run) but worth
a cleanup pass.
Merges an independent validation pass on F0/F1: a real end-to-end test
(embedded Postgres, no Docker required) driving a live SMTP session
against the exact server main.go assembles, plus fixes for three gaps
the test's own analysis found - folder/mailbox counters not updated,
mailbox quota never checked at RCPT, and the aliases table never
consulted for recipient resolution. See commits 097fa37 and 2f73898
for full details.
Per feature/3-f0-f1-validation's own gap analysis (commit 097fa37):
- InboundMessageRepository.Persist now updates folders.unread_count,
folders.total_count, and mailboxes.used_bytes in the same transaction
as the message insert (PLAN.md section 6.1 "++counters" was a no-op
before this).
- MailboxRecord carries QuotaBytes/UsedBytes; ResolveRecipient rejects
RCPT with 452 4.2.2 when a target mailbox has already reached its
quota (used_bytes >= quota_bytes), before DATA - never silently
accepting mail into a full mailbox.
- MailboxRepository.ResolveDeliveryTargets tries a direct mailbox match
first, then falls back to the aliases table (exact source_address
match takes precedence over a domain catch-all). One RCPT TO can now
fan out to multiple mailboxes; alias destinations that don't resolve
to an internal active mailbox are silently skipped, since outbound/
external forwarding doesn't exist yet. postmaster@/abuse@/dmarc@-style
aliases now actually deliver.
The E2E test's counters_updated subtest flips from a documented KNOWN
GAP (t.Skipf) to a hard assertion now that it passes, and gains two new
subtests (quota_rejected_when_mailbox_full, alias_resolves_to_destination_mailbox)
covering the other two fixes end-to-end. New unit and Postgres
integration test coverage added for all three fixes at every layer.
Verified: full suite green on Windows (unit + E2E via embedded
Postgres, no Docker needed) and via WSL Docker against the real
docker-compose Postgres (all infrastructure/postgres and diskstorage
integration tests, not just E2E). lang-lint clean.
Boots a real Postgres (no Docker), applies the real migration and seed,
starts the real go-smtp server wired as main.go does, and drives a live
SMTP session. Asserts: 550 for unknown RCPT, durable EML + row with
uid_next allocation, sha256 disk/db match, outbox event, blob dedup with
ref_count. Documents the counters gap (PLAN 6.1) as an explicit skip.
Phase: F1 validation
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Merges the first F1 sub-project: a real :25 SMTP listener for the Go
protocols service with STARTTLS, Postgres-backed recipient resolution,
content-hash-deduplicated blob storage, and transactional persistence.
See commit e501cdc for full details.
Adds a real :25 SMTP listener to the Go protocols service, giving F1's
"Nucleo SMTP" milestone its first working sub-project: application,
infrastructure, and presentation layers on top of F0's domain-only
Clean Architecture.
- ProcessInboundEmailUseCase resolves recipients against Postgres and
persists accepted mail (blob dedup by content hash, transactional
UID allocation + outbox event, never sending 250 OK before durable
commit per PLAN.md section 6.1).
- Postgres repositories via pgx/v5 (replacing lib/pq entirely).
- Local-disk blob storage with content-hash dedup, rename-before-commit
ordering so a DB row can never reference a file that isn't actually
on disk.
- A local-file TLS cert provider for STARTTLS, with an ephemeral
self-signed fallback so a missing/misconfigured cert degrades
gracefully instead of crash-looping the service (PLAN.md section 8.4).
- A github.com/emersion/go-smtp Backend/Session adapting real SMTP
traffic to the use case: unknown recipients get 550 5.1.1 before
DATA, duplicate RCPT TO for the same address is deduplicated, TLS is
hardened to the project's required PFS-only cipher/curve policy.
- A dev-seed script and Makefile target for local testing.
Explicitly out of scope (separate follow-up sub-projects): alias
resolution, SPF/DKIM/DMARC/ARC verification, Rspamd/ClamAV, Submission
ports with SASL auth, DKIM signing, outbound delivery, and the real
Traefik acme.json cert watcher.
Verified live end-to-end via WSL Docker: the full compose stack (db,
redis, migrate, auth, storage, protocols, web) reaches healthy, and a
real swaks SMTP session is accepted and durably persisted with correct
sequential UIDs, while an unknown recipient is correctly rejected with
550 5.1.1 and zero rows persisted. Built via 10 sequential tasks under
subagent-driven development, each independently reviewed, plus a final
whole-branch review that caught and fixed a TLS-crash bug, a blob
commit-ordering bug, missing TLS hardening, and missing CI integration
coverage before merge.
Known follow-up, not blocking: email_messages' partition only covers
July 2026 and needs a partition-management job before 2026-08-01.
Verified live via WSL Docker (docker compose up --wait): every other F0
service came up healthy, but web stayed unhealthy. Root cause was two
compounding issues, both invisible to static review:
1. Docker auto-sets HOSTNAME to the container ID; Next's standalone
server.js honors that env var and binds only to it instead of all
interfaces. Fixed by explicitly setting ENV HOSTNAME=0.0.0.0 in the
Dockerfile, after Docker's automatic assignment.
2. Even bound to 0.0.0.0 (IPv4), the healthcheck's `wget http://localhost`
resolved localhost to ::1 (IPv6) first inside the container, where
nothing was listening. Fixed by pointing the healthcheck at
127.0.0.1 explicitly instead of localhost.
Full stack (db, redis, migrate, auth, storage, protocols, web) now
reaches healthy end-to-end under docker compose up --wait.
Merges the F0 milestone from PLAN.md: monorepo scaffolding, initial Postgres
schema migration, the Go protocols domain layer with unit tests, health-checked
auth/storage/webmail service skeletons, local docker-compose.yaml, language
lint, and CI workflow.
Stands up the LambdaMail monorepo per PLAN.md's F0 milestone: local dev
scaffolding (.env.example, Makefile, .gitignore excluding PLAN.md/docs/
.superpowers/), the full initial Postgres schema as a golang-migrate
migration, the complete Go domain layer for services/protocols (value
objects, entities, domain events - zero external imports beyond
github.com/google/uuid) with unit tests, health-checked service skeletons
for auth/storage (TypeScript) and webmail (Next.js), a local
docker-compose.yaml wiring all services together, a language-enforcement
lint script, and a GitHub Actions CI workflow.
Built via 11 sequential tasks under subagent-driven development, each with
an independent implementer + reviewer cycle. Fixes applied during review:
DomainName.IsPunycode() added to satisfy the promised interface; ioredis
default-import typing under NodeNext; next.js bumped 15.0.3->15.5.21 to
resolve high/critical CVEs; lang-lint.mjs corrected to not flag its own
banned-terms list; web service healthcheck added so compose --wait
actually waits on it; webmail Dockerfile's missing public/ directory:
protocols go.mod version aligned with CI's Go 1.22 pin.
Known follow-ups for later phases (not blocking F0): go-arch-lint isn't
wired into CI yet; live docker compose up and migration apply have not
been exercised in this dev environment (no Docker/psql available here)
and will get their first real run in CI, which does have Docker.
