f20c942d64800cd9880a2bf1be87257400ac3280
60
Commits
| Author | SHA1 | Message | Date | |
|---|---|---|---|---|
  mixandClaude Opus 5.5
|
32f6ed626d |
findings of the security review of stages 2-6
The review of 4433688..0808779 (read-only, a separate agent) found one medium and some lows; all fixed here. Medium, proven by the reviewer: A1's userKey put the typed username into the limiter map whole, and the public forms had no body bound beyond net/http's 10 MiB for a parsed form - each refused request kept up to 10 MiB in memory for fifteen minutes. userKey now holds a 96-bit hash of the folded name, and sign-in, reset, a reset link's page and setup read at most 64 KiB (limitForm, http.MaxBytesReader). TestPublicFormsAreBoundedAndKeysShort. Lows: - Refund skipped the request that filled a bucket (count < max), so a right password at the ceiling left the shared address blocked; Refund now always undoes the Allow it follows. - The reset form did not give the address its attempt back when the name's bucket refused (sign-in did); it does. - A3's day of five links is spent only by a link that goes out, and the price of the pause - anyone knowing a username can pause that owner's reset for a day - is recorded in security.md. - longHeaderLine counted a trailing CR on a header's first line. - test.yml declares permissions: contents: read. Not changed, as the review's nits: the store matches names exactly while the limiter folds them (stricter, said in the comment now); a session from a longer window is honoured until its next Touch (the next request). go test ./internal/web/auth ./internal/milter. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> |
||
|
mixandClaude Opus 5.5
|
ec1c3a6138 |
docs: the 2026-10-17 security audit in security.md's history and the roadmap
"After the plan" of docs/security-audit-2026-10-17.md: security.md's history paragraph names A2 and the seven mediums and what closed them; the roadmap gets the row security-audit-2026-10-17 (26 of 29 plan lines; stage H is the owner's host, S4 in progress) and two proposed rows the audit left as product choices: upstream-tls-verify (M7's optional "secure") and server-dmarc-hint (S3's product side). Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> |
||
|
mixandClaude Opus 5.5
|
08087795cd |
docs, comments: the drift the audit listed (audit 2026-10-17, stage 6)
The lines under "Doc and comment drift" that the findings' own commits did not already fix: - security.md: a newer confirmation link replaces the older, a reset link is not replaced; html/template renders every page (no themes); the container keeps eight capabilities, named, not "KILL". - architecture.md: ImportKey writes over a name's key, as an import must; the error text reaches the browser on the Backup page's import refusal. - guide: with Caddy the key is in Caddy's data, so a full backup refuses until certs/ holds a copy (or back up caddy-data). - comments: sasldb2 holds application passwords in the clear (store, app/password.go); texthash folds ASCII (sender.go); the supervisor socket's group is supervisor (rpc.go); cert-reload's OpenSSL measurement was 3.0.x of bookworm; sendReset's limits (links.go). Migration 0013's comment stays as shipped. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> |
||
|
mixandClaude Opus 5.5
|
4f999f4f95 |
auth, docs: a session from a longer window is cut; the remaining nits (audit A6, A7, A10, W7)
Security audit 2026-10-17, stage 6. A10: a session created under a longer PANEL_SESSION_IDLE_DAYS kept its old expiry after the setting was lowered (the renewal arithmetic took ExpiresAt minus the new window for the last renewal). Touch now cuts such a session to now plus the window at once. TestSessionMadeUnderALongerWindowIsCut. A6, A7, W7 - security.md now says what the code does, which is what it should do: a wrong method answers 405 with Allow from the router before the role is looked at (the routes are public source); the setup token is made anew at the first /setup/ request after expiry, a stranger's included, and logged where only the operator reads it; two GETs make what must exist, the server's own DKIM key on Settings and Health. README's local trial says the compose file is the hardened deployment. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> |
||
|
mixandClaude Opus 5.5
|
5c5f49bad1 |
signer: MIME headers and Message-ID signed; a long header line logged (audit M6, M8)
Security audit 2026-10-17, M6 and M8 (nits), stage 6. M6: Content-Type, MIME-Version and Message-ID were not in h= (as in OpenDKIM's default list): a relay could change a MIME boundary or the type and the signature still verified. They are signed when present; Message-ID is not over-signed, since cleanup adds one a message lacks after the milters and an absent name in h= would refuse it. M8: a header line over smtp_line_length_limit (998) is broken by Postfix's smtp client after signing and the signature fails silently; only body lines were logged. longHeaderLine names the header, once per message, in the same words. Tests: TestSignLinesCoverTheMIMEHeadersAndLongHeaderLinesAreNamed; the h= expectations gain Message-ID; e2e dkim_shapes expects content-type: in h= (and the receiver's verification over it). Docs: security.md, architecture.md, guide; CHANGELOG. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> |
||
|
mixandClaude Opus 5.5
|
8aad5b6223 |
auth: reset links pause for a day after five (audit A3)
Security audit 2026-10-17, finding A3 (low), stage 3. sendReset spaced reset mails ten minutes apart and capped them at five an hour per account, so a stranger who repeated a username kept a live 30-minute link arriving in the administrator's mailbox five times an hour, for ever. Now five reset links a day per account (resetDayLimiter); the sixth request sends one notice instead - "Reset links paused for a day", no link, saying further requests are ignored for 24 hours and that a global administrator can set a password on Server > Users - and every later one in the window sends nothing (resetNoticeLimiter), logged either way. In memory, like every limiter: a restart clears them. Test: TestResetLinksPauseForADayAfterFive. Docs: security.md, guide § Account; CHANGELOG; plan. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> |
||
|
mixandClaude Opus 5.5
|
3a8e19b71c |
auth: sign-in limits per address and name; successes free (audit A1, A8, A9)
Security audit 2026-10-17, findings A1 (medium), A8, A9 (low), stage 3. Under the shipped Apache-on-host deployment with TRUSTED_PROXY_CIDR empty, every visitor has the Docker gateway's address, and every public limit was keyed on the address alone: ten POSTs to /login per 15 minutes (any body, successes counted) locked out every administrator; likewise /login/reset, the link pages and /setup/. IPv6 was keyed per /128. Now (internal/web/auth/limitkeys.go): - sign-in counts per address and username (10 / 15 min) and per address alone at 100; a right password refunds both (rateLimiter.Refund); - a reset request counts per address and username (5) and per address alone (50); a link's page per link, by token hash (30), and per address alone (300); setup per address as before; - an IPv6 address is keyed by its /64 (limitAddr); - proxyWatch: when the first 20 requests after a start all come from one private or loopback peer with X-Forwarded-For and no proxy is trusted, the log says so once with the TRUSTED_PROXY_CIDR value to set; - a refused sign-in or a lockout is logged, five lines per address in 15 minutes, the typed name printable and cut (refusalLog, logName). The shipped compose cannot know the gateway, so the guide's install step prints it (docker network inspect) and says to set it. Tests: TestLoginBehindOneUntrustedProxyKeepsOthersIn (the reader's case), TestLoginSuccessesDoNotCount, TestIPv6IsLimitedPerSlash64, TestProxyWatchAndRefusalsAreLogged; the existing lockout tests pass. Docs: security.md § First-run setup, § Authentication; guide § Full deployment; CHANGELOG; plan. Also the I1 plan line's Done note. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> |
||
|
mixandClaude Opus 5.5
|
1bafd78364 |
image: the TLS key's group is the panel's alone, 10997 (audit I1)
Security audit 2026-10-17, finding I1, stage 2; the help mockup is
|
||
|
mixandClaude Opus 5.5
|
e9d2cbbe28 |
findings of the security review of stage 0
The review of 2c27275..2c8e981 (read-only, a separate agent) found no high or medium; three small ones, fixed here: - reauth took a limiter attempt before the compare, right password included: with A2 asking on every user save, ten saves in fifteen minutes locked the administrator out of Account's password change, the backup and the export. A right password now gives its attempt back (rateLimiter.Refund, Module.PasswordCheckPassed; a bucket already at its limit stays blocked). TestRightAccountPasswordsDoNotUseUpTheLimit. - LoggedPath hid the tokens of /login/reset/ and /confirm/ but not of the first administrator's /setup/<token>; now it does (test case added). - The entrypoint's own `[ -e ] || : >` lines (sender_login_maps, the four inbound maps, mail.log) get the `[ -L ]` guard postfix-config.sh got for I2. Safe before too - drop_links runs first on those paths and the panel is not running yet - this is the belt and braces. The review also confirmed, with a probe of the regex against Postfix's formats, that a port-25 client can place only its own envelope in an Inbound log row, and that every path to a password, e-mail or role asks for the signed-in password. go test ./internal/web/handlers ./internal/web/auth; sh -n entrypoint.sh. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> |
||
|
mixandClaude Opus 5.5
|
b81ff9f135 |
domains: a domain's DMARC reports are deleted with it (audit A5/W2)
Security audit 2026-10-17, finding A5/W2 (low), stage 1; the mockup is
|
||
|
mixandClaude Opus 5.5
|
626bd337be |
dmarc: no custom report address on the server's own name (audit W1)
Security audit 2026-10-17, finding W1, stage 1. HandleDomainDMARC took any
address validate.Email takes; AllowedRecipients opened every stored address
on the hostname, not only the domain's hosted one; checkReportDomain
cross-checked the claimed domain only against a dmarc-reports+<tag> local
part. So a domain user of tenant A could store ops@<hostname> and from then
on anyone could mail a report claiming tenant B's domain to it, stored and
shown under B and counted toward B's 500; or store B's hosted address and
open B's mailbox without B choosing hosted. TestAllowedRecipients asserted
ops@<host> was allowed - a choice, now reversed.
Now: a custom address on the hostname is refused on the form ("choose
SelfPost hosted"); AllowedRecipients adds a domain's address only when it
is that domain's own hosted address; checkReportDomain requires a tag and
the claimed domain to equal it. The live registry (checked 2026-10-17)
holds two hosted addresses and no custom one on the hostname, so nothing
there changes.
Tests: TestCustomReportAddressOnTheHostnameIsRefused (typed, another
domain's hosted, the domain's own hosted typed as custom; an address
elsewhere still saves); TestAllowedRecipientsFollowTheDomainAddresses
inverted plus another domain's hosted address stored; untagged addresses
rejected in TestIngestMessageRejectsTagMismatch; the ingest tests deliver
to tagged addresses.
Docs: security.md § Accepted risks (DMARC), guide § DMARC reports;
CHANGELOG; plan.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
|
||
|
mixandClaude Opus 5.5
|
c50314fb4c |
inbound journal: one row per recipient, and every row counts to the cap (audit M3)
Security audit 2026-10-17, finding M3, stage 1. inboundSession.RcptTo appended every RCPT (the outbound session de-duplicates), and the end of the message inserted a row per recipient one statement at a time inside Postfix's callback; the cut to 10 000 rows every 500 counted only the log tailer's rows. Postfix accepts a repeated recipient up to smtpd_recipient_limit (1 000) and sends each to the milter, so one message was a thousand rows, uncut until the six-hourly prune. Now: the inbound milter folds a repeated recipient (case-insensitive, as cleanup's been_here() with BH_FLAG_FOLD does, and as session.RcptTo does); a message's rows go in one transaction (store.InsertInboundSpamMessage, which replaces InsertInboundSpam in milter.InboundStore); the count and the cut moved from the tailer into the store (countInboundWrites, an atomic counter on Store) so the milter's rows, the tailer's and the tailer's start-up re-reads all count. store.InboundSpamLogMaxRows is the one cap. Tests: TestInboundRecipientRepeatedIsOneRow (1 000 RCPT in two cases, one row); TestInboundJournalIsCutWhicheverWriterWrites (a small cap: a message's rows and refusals together cross the mark, the table is cut); the tailer's own counting test went with its counter. go test ./internal/store ./internal/milter ./internal/logtail (Windows-only Follow failures aside). Also gofmt of a test line from the M2 commit. Docs: security.md § Mail path, architecture.md; CHANGELOG; plan. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> |
||
|
mixandClaude Opus 5.5
|
d254a9952d |
milter: the level-2 domain limit is the login's domain's (audit M1)
Security audit 2026-10-17, finding M1, stage 1. overLimit took domainOf(MAIL FROM) whatever the login, and MailFrom ran it with an empty login: with smtpd_delay_reject at its default Postfix runs no sender restriction at MAIL and sends the milter's MAIL event at once, so a stranger on 465/587 with no AUTH, or another tenant's application, could name a victim domain in MAIL FROM and take its slots for the 10-minute in-flight TTL, get the victim's application answered 451, and have a limit row (and the "a rate limit is refusing mail" notification) written under the victim's domain. Now overLimit returns at once without a login and takes the domain from ApplicationByLogin (loginDomain, fail-open like the rest); recordRejected's Domain is the login's too. The send log already counts a domain's messages by the login's domain (store.CountMessages, loginDomainID), so the limit and the count now agree. The null-sender special case is gone: every message is held to its login's domain. Tests: TestRateLimitIsNotReservedByAnotherSendersMailFrom (a stranger with no login; another tenant's application) - fails on the old code at the second MAIL FROM; the domain-limit tests sign in as app1 at example.com. e2e domain_limit_not_reserved_by_a_stranger: a domain with a limit of one, three unauthenticated MAIL FROM naming it on 465, then its own message passes and its second is refused 451. Docs: security.md § Mail path; CHANGELOG; plan. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> |
||
|
mixandClaude Opus 5.5
|
da2e8b2ffe |
milter: the sender check keeps two From: and two Sender: values (audit M5)
Security audit 2026-10-17, finding M5 (low), stage 0. senderSession.Header appended every From:/Sender: value; Postfix bounds one header at 60 000 bytes but not their number, and the signer's 1 MiB cap runs after the sender check, so an application could make it hold megabytes per connection (the reader's test: a hundred lines, 6 MB). judge needs only none / one / more, so two of each are kept; the refusal text joins what is kept, as before. Tests: "a hundred From: lines" in the judgement table (refused, more than one From: line); TestSenderKeepsTwoFromAndSenderLines. go test ./internal/milter. Docs: security.md (the sentence "a few headers each"); CHANGELOG; plan. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> |
||
|
mixandClaude Opus 5.5
|
683fdafc6d |
logtail: Inbound log takes Postfix's own tail, not a recipient's (audit M2)
Security audit 2026-10-17, finding M2, stage 0 - proven on the live server: Postfix quotes a refused recipient in its reply in internal form (quotes removed, "; < > =" kept: extract_addr, tok822_internalize), so RCPT TO:<"; from=<ceo@…> to=<board@…> proto=ESMTP helo=<x>; "@domain> spelled a whole tail inside the reply, and inboundRejectRe, whose reply ended at the FIRST "; from=<", took the stranger's sender, recipient and inbound domain for the row. Now the reply is greedy and the tail anchored: "; from=<…> to=<…> proto=\S+", a HELO name holding none of Postfix's NEUTER_CHARACTERS (smtpd.c:1634, helo_cmd neuters ` <>()\";@`), and one-word SASL fields to the end of the line (a `.*` SASL tail would have let a quoted MAIL FROM spell the whole tail). Belt and braces: a reply that still holds "; from=<" - a line with two tails - makes no row, so neither a recipient nor an un-neutered HELO name can choose one. The quote/bracket/space filter on from/to stays. Tests: TestInboundRecipientCannotWriteTheRow (the live line's shape, and with a forged MAIL FROM too); TestInboundHeloCannotWriteTheRow now feeds HELO names as Postfix logs them (neutered) and expects no row for the un-neutered ones. go test ./internal/logtail on Windows (the three known Windows-only Follow failures aside). Docs: security.md § Mail path; CHANGELOG; plan line ticked. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> |
||
|
mixandClaude Opus 5.5
|
a37aba6594 |
user-reauth: the user form asks for your account password (audit A2)
Security audit 2026-10-17, finding A2 (high), stage 0; the mockup is
|
||
|
mixandClaude Opus 5.5
|
b8e156c9f4 |
password-reset: a confirmed e-mail, then a link to reset the password
Agreed by the owner on 2026-10-13 ("the e-mail must be confirmed");
notifications to an unconfirmed address settled on 2026-10-17: "Без
подтверждения слать". Mockups
|
||
|
mixandClaude Opus 5.5
|
be98efa549 |
docs-audit: architecture, schema, security and the rest against the code
The owner's "проверь, что документация актуальна с кодом" (2026-10-16).
Read-only reviews checked every present-tense claim against the code;
each finding was read in the source before it was fixed here.
architecture.md: held inbound mail has its row from the journal milter,
not the tailer; the antispam action is accept or tempfail, anything else
stops the start; tls-internal's certificate is 0644; the signer logs to
stderr; the log's fingerprint is the inode, not the device; a mail.log
postlogd makes is 0640 (maillog_file_permissions); an unreadable stored
place starts at end-of-file too; the test-message route; no map in page
data; design-first compares whitespace aside; the queue and sasldb2.new
are not backed up; antispam/ under /data; POSTFIX_QUEUE_DIR's desync;
SELFPOST_BACKUP_PASSWORD (the CLI's).
schema-migrations.md: inbound_spam_log's upstream columns (v12),
send_log's sender reason, row 9's client_ip.
security.md: /data/opendkim since 2.9.0 (removed, or the start refused),
sasldb2 under /data/sasl, the trigger for mandatory encryption, the
password minimum counted in characters.
development.md: Staticcheck in CI, the e2e ports 20025 and 20053/udp.
README: the client IP list refuses at MAIL FROM, overrides are per
application. product.md: the optional rspamd sidecar.
Code: validate.AdminPassword counted bytes, so six Cyrillic letters
passed for twelve characters; it counts runes, as the backup password
always did (TestAdminPasswordCountsCharacters). envdoc reads
cmd/selfpost-backup too, and a name kept in a constant. A stale comment
of cmd/panel's run ("three roles").
go vet + go test (26 packages) and make e2e on the test server for this
tree.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
|
||
|
mixandClaude Opus 5.5
|
f2ff8b4fea |
test-email: a test message from Server › Health
Decided by the owner on 2026-10-13 (out of preflight); the sender is his
choice of 2026-10-16 ("Да, адрес годится"); mockups
|
||
|
mixandClaude Opus 5.5
|
503939c6cf |
origin-strict: a change with neither Sec-Fetch-Site nor Origin is refused
Agreed with the owner on 2026-10-13 (in place of csrf-tokens), taken up on
2026-10-16 ("Да"). originAllowed (internal/web/security.go) returns false
in the "neither header" branch: such a POST is answered 403 and logged
like a cross-origin one. It was let through until now as an accepted
risk, which left a browser that sends neither open to a request forged by
any site. A browser that follows Fetch § 3.2 never sends a POST without
Origin; one too old to (Firefox before 70) can no longer save anything.
What posts to the panel without a browser: the e2e suite's client, which
now sends Origin on every form (panelClient.postForm); cmd/panel's tests
already did. Nothing else posts to it (the CLI runs in the container).
Tests: security_test's case of no headers is refused; e2e posts with
neither header and gets 403 (testFormPostWithTheBrowsersOrigin). go vet +
go test (26 packages) and make e2e on the test server for this tree.
Docs: security.md - the accepted risk struck, the ADR's decision and
trade-off say refused, the origin-check bullet; architecture § Web,
guide § the proxy; CHANGELOG [Unreleased] § Security; roadmap row done.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
|
||
|
mixandClaude Opus 5.5
|
49c872f836 |
backup-reauth: the account password before a backup or an export
Agreed with the owner on 2026-10-13, taken up on 2026-10-16 ("Да"); the
mockup is
|
||
|
mixandClaude Opus 5.5
|
22ce1bdfbb |
postfix: long, non-repeating queue ids
The owner, 2026-10-16, of three ways to stop a reused queue id reaching an older row (found by the technical review of inbound-upstream-result, L6): "1" - enable_long_queue_ids=yes. A short id is the microsecond of arrival and the queue file's inode in hex (Postfix 3.10.13, src/global/mail_queue.h, MQID_SH_*) and repeats once the inode comes round; a later message with no send-log row of its own (port-25 mail, a bounce Postfix sends) could then rewrite an older Outbound log row of the same id and recipient, its page could show the later message's stored lines (security.md's known exposure, "enable_long_queue_ids ... ends it"), and the inbound journal's rows needed detaching. A long id adds the time in seconds and does not repeat (MQID_LG_*; proto/postconf.proto: existing queue files keep their names, postqueue -p gets a wider Queue ID column). Every pattern of the panel takes both shapes (queueLineRe's long branch, the [0-9A-Za-z]+ of the others, queueEntryRe) - checked one by one. Tests: e2e reads a long id on an outbound send-log row (testExpiredMessageIsClosed) and on an inbound journal row (testInboundUpstreamDown). go vet + go test (26 packages) and make e2e on the test server for this tree. Docs: security.md (the collision bullet closed for new mail, the upstream bullet), architecture § Log tailer, delivery-log-storage's note, the ForgetInboundQueueID comment; CHANGELOG [Unreleased] § Fixed. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> |
||
|
mixandClaude Opus 5.5
|
d51549385b |
inbound-upstream-result: the upstream's answer in the Inbound log
test / test (push) Canceled after 0s
The owner, 2026-10-16: the Inbound log lacks the upstream's answer - "Не понятна судьба письма: принято, отклонено, хост не доступен"; the result in the row only, a page per message later if needed. Mockups |
||
|
mixandClaude Opus 5.5
|
9420d8c03d |
entrypoint: /data/opendkim removed, as planned for the release after 2.8.0
2.8.x copied the keys from /data/opendkim to /data/dkim and kept the old tree for a rollback to 2.7.x "until the next release" (docs/plans/panel-dkim-signing.md, Q2); the owner's "потом режь 2.9.0" (2026-10-16) makes this that release. The entrypoint now removes the tree once /data/dkim holds a key, and says so. A /data with keys under the old tree only - one that never ran 2.8.x - stops the start with a message to run 2.8.x first, rather than let the panel give the server's own name a new key its published record would not match; the tree is left as it was. Gone with it: the copy into /data/dkim.new, the tree's normalisation, its exemption from the general hand-over, its key link in drop_links, and the backup's exclusion of "opendkim" (dead: the panel never sees the tree now). Tests: e2e TestStartsOnA28Data replaces TestStartsOnA27Data - a /data as 2.8.x left it (both trees, a queue of foreign uids, a rotation due): the old tree is gone, the key under /data/dkim is the one left, the selector table names it, the queue is Postfix's by name; TestRefusesA27Data - the start stops, says why, leaves the old key and makes no new one. go vet + go test (26 packages) and make e2e (all pass) on the test server for this tree. Docs: guide § Upgrading "To 2.9.0" (rollback to 2.8.x needs nothing, to 2.7.x only from a 2.7.x backup; the 2.7.x rollback paragraph says it is for 2.8.x only and keeps naming 2.7.0), guide § backup, architecture (the old tree, § Persistence's table and the restore line), security § Backup, development's e2e paragraph, the plan's key step; CHANGELOG [Unreleased] § Removed. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> |
||
|
mixandClaude Opus 5.5
|
f1d90a2a1b |
go-milter: build it patched not to keep every header
test / test (push) Canceled after 0s
The security documentation's last open item of panel-dkim-signing (the owner's decision, 2026-10-16): go-milter v0.4.1 added every header of a session to a map of its own, emptied only on ABORT, and handed it to Headers - which no milter of the panel reads. Postfix bounds one header, not their number, so a message of nothing but headers up to the message size limit was held whole by the library in the sender check, the signer and the journal milter on 465/587 and in the inbound journal milter on port 25, where strangers send; the signer's 1 MiB refusal came after it was all in. third_party/go-milter is v0.4.1 from the module cache (BSD 2-Clause, LICENSE kept) with session.go changed: CodeHeader hands the header to Header and keeps nothing, CodeEOH hands Headers an empty map. PATCHES.md says what and why, and how to take a new upstream version. go.mod replaces the module with it; the Dockerfile copies third_party before "go mod download". TestTheLibraryKeepsNoHeaders holds it through the library's own client over two messages of one connection. Comments, security.md (the accepted risk becomes a note), development.md's library table, the plan and CHANGELOG follow. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> |
||
|
mixandClaude Opus 5.5
|
0cceb549cc |
sender check: a Sender: line is held to the rule From: is held to
test / test (push) Canceled after 0s
The owner's decision (2026-10-16) on the security review's F1: a Sender: line - the sending agent's, shown beside From: as "on behalf of", and signed since 2.8.0 - was neither held to the login nor, before, signed: an application could write its own From: and Sender: ceo@bank.com under this server's signature. Now, when a message has one, it must be one address the application may send as, and at most one such line (RFC 5322 § 3.6); From: is judged first. The refusal is "550 5.7.1 Sender: <address> - <why>", a send-log row of reason "sender" (store.ReasonSender), tagged "wrong Sender"; the message page names the line. The log tailer recovers such a row from cleanup's line as it does a From: refusal (fromRefusalRe reads either). Help text as in the design commit d9d9af3; guide, architecture, security and CHANGELOG say it, security.md keeps Resent-From: (signed, not judged) as an accepted risk. Tests: TestSenderLineHeldToTheSameRule (own address, a name, another application's, a stranger's, two addresses, two lines, From: first), the recovery of a Sender: refusal line, the tag, the message page; e2e sender_check_holds_the_sender_line (own Sender: passes and is signed, another domain's refused at DATA and logged). Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> |
||
|
mixandClaude Opus 5.5
|
2cfd239d16 |
docs: panel-dkim-signing stage 3 - the documents without OpenDKIM
test / test (push) Canceled after 0s
architecture § OpenDKIM becomes § DKIM signing (keys, the selector table and when it is rewritten, the two sockets, what is signed and the limits, the old tree); the process table, start order, liveness, the mail-path diagram, § Sender check, Persistence and the restore Resync follow the code. security.md: the panel signs with the login's key, the keys are the panel's alone, group selfpost and CAP_KILL as they are now; OpenDKIM's accepted risks go; new ones: what sendmail(1) submits is signed as the server, Sender: is signed but not judged (the owner's decision), and go-milter keeps every header per session, so the signer's 1 MiB refusal does not bound the library's copy. guide: Health, the container log's dkim-signer lines, backups, the long-line limit, and upgrading to 2.8.0 - with the rollback to 2.7.x: its start reads the queue as its own postfix (uid 101) before it hands the queue over, so the hand-over by name is run in the 2.7.x image first (measured on a copy of the test server's queue: postsuper failed before, passed after). development, README, SECURITY, product, the assumptions register's header: what left the image. entrypoint.sh: a comment, the rollback note as the guide gives it. Plan stage 3 lines ticked, roadmap 12/14. Written by four agents, one document each, against a brief of what the code does; reviewed here. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> |
||
|
mixandClaude Opus 5.5
|
a3050afb66 |
base-trixie: findings of the reviews
test / test (push) Canceled after 0s
Security review ( |
||
|
mixandClaude Opus 5.5
|
1c6e04b7d3 |
base-trixie: the image on Debian trixie, Postfix 3.10.13; Received: hidden on 465/587
test / test (push) Canceled after 0s
The runtime base is debian:trixie-slim, the build stage golang:1.26-trixie:
Postfix 3.10.13 with upstream's 2026 security fixes up to the August
batch, libmilter 8.18.1, OpenSSL 3.5. Every citation of another program in
code, scripts and documents was re-read against the trixie sources
(dev/upstream; bookworm's kept in dev/upstream-bookworm) and updated;
development.md's cited-package table holds the trixie versions.
What the re-reading changed in behaviour:
- logtail: reject lines with sasl_method/sasl_username/sasl_sender
(smtpd_check.c:1019-1026); cleanup's "removed (canceled)" forgets a
port-25 id (cleanup_api.c:363-364).
- internal/postfix/retry.go reads postconf's stdout only: since 3.9 a
deprecated-parameter warning on stderr made one value too many and the
panel fell back to the compiled-in retry policy.
- logrotate-loop.sh: logrotate 3.22 exits 1 when it ignores a config;
the loop refuses on that output whatever the status.
- postfix-config.sh: cyrus_sasl_config_path set explicitly (bookworm's
07_sasl_config.diff is gone); maillog_file_permissions=0640; mynetworks
explicit; Dockerfile: SYNC_CHROOT= in /etc/default/postfix,
opendkim-tools dropped, openssl named, systemd-standalone-sysusers so
cron does not bring the whole of systemd.
- The selfpost group is 10998, outside the system range 100-999: in
trixie a package took 998 and the build failed (owner's decision:
a conflict-free id, the distribution's own ids untouched). The
entrypoint hands /data back by name when the user or the group differs.
Guide, backup refusal text, help (design commit
|
||
|
mixandClaude Opus 5.5
|
aaa94120be |
port 25 ends DATA only at CRLF.CRLF: smtpd_forbid_bare_newline=normalize (port25-bare-newline)
Postfix 3.7.11 leaves its 2023 SMTP-smuggling defence off (default "no" before 3.9, src/smtpd/smtpd.c:807), and the inbound relay's smtp/inet listener did not turn it on: a "." line ended by a bare LF ended DATA and what followed was read as the next transaction. The listener now sets normalize (bare_lf_mask_table, smtpd.c:1656; the DATA loop drops such a "." line, 3649-3652), the setting Postfix recommends and 3.9's default. The submission ports keep the default (own-mta-builds, section 5). Tests: e2e inbound_bare_newline_is_one_message sends the smuggling sequence over raw SMTP (net/smtp's DotWriter would rewrite the bare LF) and finds one message at the upstream with the smuggled part inside; port25 config check asserts the parameter on smtp/inet. On the server: make e2e exit 0 (511 s), go vet + go test 27 packages ok. Docs: CHANGELOG [Unreleased] Security, security.md, architecture.md, guide.md; roadmap: port25-bare-newline, base-trixie and panel-dkim-signing agreed by the owner in this order. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> |
||
|
mixandClaude Opus 5.5
|
a39eda5a0b |
audit 5: findings of the two reviews
- The e2e's numeric-settings test ran a container that passed every gate
and never ended: the good values are checked beside a bad one after
them; each run is named, bounded to a minute and removed.
- The gates refuse a value under the parameter's minimum (a zero window
kills anvil, a zero milter timeout smtpd) and one past an int, in
seconds for a time value (src/global/conv_time.c, src/anvil/anvil.c).
- OpenDKIM waits, five seconds at most, for the relay's socket.
- The relay marks its lines `syslog: `, writes control characters out,
reads into a queue at once and drops past 200 lines a second, saying
how many (security F1, F2; a stuck relay would hold OpenDKIM in
syslog(3)).
- smtp_tls_mandatory_protocols >=TLSv1.2 too.
- extract-cert.sh keeps the old files' group and sets its trap before
mktemp.
- crashexit spares cert-reload; architecture.md "four times running" and
what else exits the container; security.md the restart after five panel
exits and the /dev/log socket.
- Measured against the technical review: smtpd has the selfpost group
(Groups 104 998), so it did reach the supervisor socket before this
stage.
Accepted, reasons in the plan: the logrotate state file in the postfix
user's directory, envdoc_test's reading of `${VAR:-}` only, crashexit's
window pruning untested.
Tested on the test server: go vet + go test of 27 packages, make e2e (all
pass, 592 s) on this tree.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
|
||
|
mixandClaude Opus 5.5
|
dab7715e80 |
audit 5: the supervisor socket the panel's alone, OpenDKIM logs and oversigns From
test / test (push) Canceled after 0s
M17: the supervisord control socket is root:supervisor 0770, a group whose one member is the panel; OpenDKIM and Postfix keep `selfpost` for the files under /data and no longer reach the socket. L30, read in OpenDKIM 2.11.0~beta2's source first: it logs through syslog(3) only and has no Logfile option, and nothing in the container served /dev/log. internal/syslogrelay (the panel binary with -syslog-relay=/run/syslog/log, program syslog-relay before OpenDKIM, /dev/log a link to its socket) passes the datagrams to stdout, error and worse marked. opendkim.conf: Syslog yes, Canonicalization relaxed/simple, OversignHeaders From (RFC 6376 § 3.4, § 8.15). e2e: OpenDKIM's start line reaches the container log; every program RUNNING; DKIM verifies. Tested on the test server: go vet + go test of 26 packages, make e2e (all pass, 626 s) on this tree. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> |
||
|
mixandClaude Opus 5.5
|
eaa7de0754 |
audit 4: findings of the two reviews
test / test (push) Canceled after 0s
- Two inbound domains on one upstream could never change its TLS mode,
each save refused for the other (medium): SetTransportAs changes the
mode for every domain on the upstream when the one saving may change
them all; a domain user who may not is refused with the mode, not the
other tenant's domain name (security F1: the first version named it).
- A deletion's leftover DKIM key is removed by Resync (Reload
configuration, a restore's first start) for names no domain has.
- The backup's write deadline moves before Check.
- hasUnsavedEdit skips password and file fields and compares a multiple
select's options; htmx.config.timeout 30 s, so a refresh never answered
ends and is tried again; a failed refresh's promise is handled.
- An empty filtered Outbound log says no message matches the filters.
- The dummy bcrypt hash is made at start; a DMARC part over 8 MiB is
refused for its size like the other two limits.
- Health's Configuration box says what Reload writes (design first,
|
||
|
mixandClaude Opus 5.5
|
07be0cfb58 |
audit 4: sign-in timing and sessions, server timeouts, polling that recovers
test / test (push) Canceled after 0s
Steps 1-3 of stage 4, tested together (their docs share files).
Step 1 (M8, L15, L16): an unknown name pays a bcrypt comparison at the
panel's work factor too (noSuchUserHash, made once) -
TestLoginTakesAsLongForAnUnknownName holds the medians within a factor of
two; the Account form's current-password check has its own limiter, per
user (AllowPasswordChange); a sign-in destroys every session the request
presents before it issues its own.
Step 2 (L14): newHTTPServer - ReadTimeout 30 s, WriteTimeout 60 s,
IdleTimeout 2 min beside ReadHeaderTimeout; the full backup moves its write
deadline to two hours through http.ResponseController (e2e: the log shows
no failure to move it).
Step 3 (M11), read in htmx 2.0.4's source (dev/upstream/htmx-2.0.4, the npm
build byte-equal to the shipped htmx.min.js): the poll passes source: box,
and the next refresh is scheduled from htmx:afterRequest, which htmx fires
on load, error, abort and timeout, and after a swap on the nearest parent
left (the box found by detail.target's id); failures in a row double the
wait to 60 s. RequireAuth answers htmx with 401 + HX-Redirect: /login,
which htmx follows before it looks at the status. testdata/poll_test.js
runs panel.js in node against a stand-in DOM and clock (TestPanelJS; skipped
without node, which the test server lacks - run here, and it fails on the
old panel.js).
Also measured: the stage-3 e2e regression for H1 fails as it should on HEAD
with the old entrypoint ("check data/postfix/e2e_extra_map: permission
denied", the panel restarting).
Tested on the test server: go vet + go test of 25 packages, make e2e (all
pass, 639 s) on this tree.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
|
||
|
mixandClaude Opus 5.5
|
242515c305 |
audit 3: backup CLI, envelope budget, directory sync, docs; findings of the reviews
Steps 3-6 of stage 3 and what the two reviews found in steps 1-6, in one commit: the review fixes sit in the same hunks. Step 3, CLI backup (M3, M5, L1, L2; L4 of the stage's list): - backup.snapshotDB refuses a database that is not there and opens the source mode=ro (measured: mode=ro creates no file). - selfpost-backup holds an encrypting password to validate.MinSecretFilePasswordLen; -o of both directions writes a temporary file beside the name and renames it whole; -decrypt peeks the first chunk (password and type authenticated) before the output exists. - what a backup leaves out (a link under data/, a link under certs/ to a directory, a socket...) is told to OnWarn; certs/ with no file is backup.ErrNoCerts, refused by Check and Create, named by the panel. - parseTrustedProxies: a bare address takes /128 when written with a colon (::ffff:10.0.0.1/32 was ::/32). Step 4, secretfile (M4, L3): scrypt's V at most 128 MiB, r <= 64, p <= 2; bytes after the last chunk are ErrCorrupt. Step 5 (L7): atomicfile.SyncDir after the rename, used by Write, replaceSASLDB and the CLI; atomicfile_test.go. Step 6: security.md and guide on backup and restore. Reviews (technical, security): - H1: a restore of a server with DMARC on was refused on every start - the archive records no owner, tar as root left the DMARC maps root's, and the entrypoint only chowned the maps it named. It now hands every file beside queue/ in /data/postfix to panel:selfpost 0640; e2e writes an extra map before the backup and finds it panel:selfpost 640 after a root unpack. - the comparison reads regular files only (Lstat, size, LimitReader); a record unpacked again wins over the earlier check's mark; ConsumeRestore removes both; the mark is synced; a journal file beside a freshly unpacked database is refused. - AttachRefusalLine refuses an empty line as the old statement did; skipped entries named in words; -decrypt of an empty payload is no failure; guide wording on the shell's '>' and a killed CLI; the in-place restore's 'mv .' (older than this stage) fixed. Tested on the test server: go vet + go test of 25 packages, make e2e (all pass, 643 s) on this tree. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> |
||
|
mixandClaude Opus 5.5
|
9eddd49602 |
audit 2c: findings of the two reviews
test / test (push) Canceled after 0s
Technical review (Opus, 11 findings, 2 medium) and security review (Fable, no high or medium) of 3c86280..HEAD. Both confirm the port gives dkim_mail_parse()'s result on every input it parses, and that no line the check passes is signed with another domain's key. - F1 (medium): OpenDKIM's parser takes lines no mail program reads as one address - "ceo@bank.com:me@mine.com", "Name <ceo@bank.com:me@mine.com>", "<b@evil.com, a@mine.com>", "ceo@bank.com <me@mine.com>", "ceo@bank.com@mine.com" - and signs for the last while a lenient reader shows the first. The line must now also be one address in RFC 5322's grammar (net/mail, display names in any charset passed through), at the domain OpenDKIM signs for. A quoted display name may still hold anything. - F2 (medium): the bindings were asked about OpenDKIM's form of the address, quotes and all, so "x"@mine.com passed for an application holding @mine.com while x@mine.com belonged to another. They are now asked about the reader's address, its local part unquoted, as Postfix holds the envelope's. - F3 / security (low): Bindings.Allowed and DomainOf fold case in ASCII only: Postfix folds as Unicode (casefold.c, ICU), but every key and login the panel writes is ASCII, and the Unicode fold let the Kelvin sign pass for "k" (net/mail then refuses such a line on the domain test). - F4 (low): a TERM between "postfix start-fg &" and "child=$!" is no longer lost: the wrapper stops Postfix as soon as the pid is known. - F5 (low): for startsecs after a start any failed reload passed as "not running". postfix-reload is now build/postfix-reload.sh: exit 75 when master -t says no master runs, else exec postfix reload; RunOnce returns EndError with the status, and only 75 is taken as done. supervisor.State goes. - F6 (low): crashexit spares postfix-reload too - a reload that could not be spawned is reported to whoever asked, not a reason to stop the container. - F7 (low): a certs/ that is itself a link is resolved before the walk (filepath.WalkDir does not follow its root). - F8 (low): at shutdown the tailer reads what is left of the rotated file before saving its place. Not changed: a refusal line still waiting on the old file at shutdown is not kept - the panel writes refused rows itself, only a kill loses one, and that window is the minute's grace. - F9 (low): pct read as text - empty or unreadable is 100, the rest held to 0..100. - F10 (low): the directory is synced after sasldb2's rename. - F11 (low): POSTFIX_STARTSECS by expr, so "010" or "08" in MILTER_WAIT_TIMEOUT are decimal, not octal or an error. - Security (low): pagehide also empties defaultValue - the input's value attribute and the textarea's text - not only .value. A sasldb2.new left by a killed panel is removed at start (entrypoint) and never archived. security.md says that links under certs/ are followed and by whom they can be placed. Tests: sender check cases for F1 (four lines), F2 (a bound address quoted, for its owner and for another), F3; EndError's state and status; pct cases. architecture.md and CHANGELOG describe the reader rule. Stand (2026-10-15, make e2e, exit 0); go vet and go test ./... pass on the server (24 packages). Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> |
||
|
mixandClaude Opus 5.5
|
3584adcbc2 |
sender check: From: read with OpenDKIM's own parser; L8 closed
test / test (push) Canceled after 0s
Audit 2026-10-12, stage 2c, the owner's decision of 2026-10-14: the check
decides by the address OpenDKIM signs for, whatever a display name or a
comment holds.
internal/milter/mailparse.go ports libopendkim's dkim_mail_parse() with
dkim_mail_first_special(), dkim_mail_token() and
dkim_mail_matching_paren() from Debian's opendkim 2.11.0~beta2-8+deb12u1,
line for line: the '<' form, group and list punctuation, an '@' before the
first ':', ';' or ',' deciding where the address starts, source routes
dropped, quoted strings and domain literals kept with their delimiters,
comments skipped, whitespace between atoms one space, no encoded word
decoded; and the C's own quirks ("@d" gives d as the local part too).
testdata/mailparse_harness.c compiles the C from the image's source; the
test feeds it and the port 200 019 lines (the hand-written cases and
200 000 random ones over the specials) - 93 443 parse and agree byte for
byte, 106 576 fail and agree on the error (on an error the C leaves its
in-place result unterminated, and OpenDKIM uses neither part then).
judge(): one From: field and one address stay required (moreAfter: only
list or group punctuation may follow); what OpenDKIM cannot read is
refused (mlfi_eoh() would not sign it); the address is lower-cased in
ASCII only, as dkimf_lowercase() and Postfix's tables do. Found reading
dkimf_apply_signtable(): OpenDKIM matches its refile SigningTable against
"user@domain" printed into MAXADDRESS + 1 = 257 bytes, so a longer address
is matched cut short - a quoted local part padded to end there in
"@another.tenant" would be signed with that tenant's key while its real
domain is the application's. Such an address (over 256 bytes) is refused;
one that fits is matched whole. The display-name and comment scan, the
"one @ outside quotes" rule and the SigningTable reader (Signers) go; the
check asks for header values with their leading space as OpenDKIM does
(OptHeaderLeadingSpace; Postfix keeps a v2 milter's event mask, milter8.c
milter8_negotiate()).
L8 closed: refused or passed no longer depends on other domains of the
server; security.md's accepted risk is struck. The shared other-sender
lines are now held to "never signed for the other sender": unit test (a
passed line is read at the application's domain) and e2e (a passed line
arrives signed for the application's domain, none for the other).
Documents: architecture "How the header is read", security's sender
check paragraph, the guide; the help topic per
|
||
|
mixandClaude Opus 5.5
|
cad1182150 |
audit 2c: what only the stand could answer, measured and written down
test / test (push) Canceled after 0s
All 2026-10-15; each written beside the code or text it settles. - A-43: /usr/sbin/postlog is -rwxr-xr-x root:root in the image (postfix 3.7.11-0+deb12u1; dpkg-statoverride lists postdrop and postqueue only): Debian does not install it set-gid, so it is no use to dmarc-ingest (cmd/panel/dmarc_ingest.go). - C-X1: OpenSSL 3.0.22 in the image prints the same public key from `x509 -pubkey` and `pkey -pubout` for RSA-2048, P-256, P-384 and a SEC1 P-256 key - cert-reload's same_pair holds for certbot's ECDSA keys. - C-K5, D-guide:933/:233: on the live server (Docker 29.6.1, userland proxy on 0.0.0.0 and ::, no IPv6 on the network) an outside IPv4 client and the host's own public IPv4 address are logged as themselves; the host's IPv6 address and 127.0.0.1 arrive as 172.18.0.1. Over IPv6 every client is the gateway: one level-1 budget, an application IP list cannot name them, rspamd sees the gateway. Guide paragraph under level 1, notes at the milter and allow-list, comments at the anvil setup and authip.go, and a Security entry in the CHANGELOG (an allow-list naming the gateway would admit every IPv6 client). - D-guide:61: a published port nothing listens on (a throwaway container, -p 20999:25) is refused from outside over IPv4 and from the host's public IPv4, accepted and closed through 127.0.0.1 - Docker's proxy, which IPv6 goes through too (not probed over IPv6 for this). - C-M8: the live installation has received one DMARC report, of one record: nothing to set maxReportRecords by yet - said at the constant. - D-S7, D-arch:516, D-arch:532: a new database filled through the store to 120 days at 1 900 messages a day (228 000 messages, 1.37 million lines, 464 MB) on the test server (2 cores): the longest write is the line prune after a 30-day gap, 1.9 s for 342 000 lines (0.4 s with nothing to delete); the send-log delete 0.6 s; 2 000 lines of a tailer pass 18 ms; VACUUM INTO 1.7 s; a second connection writing every 50 ms waited 1.8 s at most, no SQLITE_BUSY - inside busy_timeout's 5 s. architecture.md's "about a second per million lines" was five times too fast; its "six to ten lines of 200 bytes" is 6.0 lines of 146 bytes on the live data. - D-B16: which favicon is fetched, by the probe server's log: Edge 155 the SVG, the desktop app's embedded Chromium 152 the PNG, with or without sizes="any" - layout.html's "vector wins where supported, raster catches the rest" holds; Firefox and Safari not measured. - D-arch:121: the allow/deny trial again, in rspamd/rspamd:3.11.1-3 with today's multimap.conf: every mix of allow and deny entries matching one message ended reject; a changed map was used about 10 s later. - D-sec:424: rspamd runs as _rspamd, uid 11333; the panel is uid 999. Stand: go vet and go test ./... pass on the server (24 packages); make e2e exit 0. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> |
||
|
mixandClaude Opus 5.5
|
fbd0acd035 |
audit 2c: the rest the sources contradict, and the documents
test / test (push) Canceled after 0s
Code and config: - C-M7: a DMARC report's explicit <pct>0</pct> stays 0; only an absent pct takes RFC 7489 § 6.3's default 100 (Pct *int). - C-D1: docs/development.md now tables the Debian versions the citations were read at (postfix 3.7.11-0+deb12u1, opendkim and libopendkim11 2.11.0~beta2-8+deb12u1, libmilter1.0.1 8.17.1.9-2+deb12u2, libsasl2-2 and sasl2-bin 2.1.28+dfsg-10, supervisor 4.2.5-1, logrotate 3.21.0-1 - read off the image with dpkg-query, 2026-10-15); the e2e stand's new checkCitedPackageVersions fails when the image it built differs. - C-E2: the e2e zone's SOA serial only grows (CoreDNS v1.14.6's file plugin reloads on a grown serial, not on mtime); Corefile comment says so; TestZoneSerialGrows. - C-K8, C-K9: the e2e compose and stage comments say how --project-directory and .env resolve. - D-arch:154: postfix-config.sh's "Postfix insists" comment says what TLS_README advises and that smtpd reads the key as root. - B-G7: inflight.go's reservationTTL comment: a dropped client does give ABORT (smtpd_proto() -> mail_reset() -> milter_abort()); the TTL is for smtpd dying or the milter connection breaking. - D-guide:301: the nginx fragment's first-certificate command now reaches certbot (--entrypoint certbot, --standalone, port 80 free), and the renewal loop passes --webroot -w /var/www/certbot - a standalone certificate would otherwise renew standalone while nginx holds port 80 (certbot docs, "Renewing certificates": the issuing plugin is used "unless you specify other plugins or options", and options of a successful renewal are saved). The register's proposal missed that. Documents: every D-arch, D-guide and D-sec claim the register's Answers E found false or partly true, rewritten - by three Sonnet agents, one per document, then reviewed: two proposals were obsolete after this stage's fixes and not used (a bounce's From: line - extract_from = "smtp" now; "Origin: null" - Referrer-Policy is same-origin now); a stand observation the agent had dated from a commit is now tied to the e2e test that checks it; the SQLite writers paragraph describes _txlock=immediate; the log tailer's rotation paragraph describes the minute's grace; the guide's nginx steps were put in working order (links and modes before `up`, no reconfigure step). "Deferred" is kept for Postfix's queue in the guide too. guide.md's CRLF from the agent normalised. Stand (2026-10-15, make e2e, exit 0, the version gate and zone serial test included); go vet and go test ./... pass on the server (24 packages). Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> |
||
|
mixandClaude Opus 5.5
|
31c2b50457 |
antispam: rspamd strips a sender's own verdict for every client
test / test (push) Canceled after 0s
rspamd 3.11.1's milter_headers skips its verdict routines - removes none of the sender's X-Spamd-Result / X-Rspamd-Action, adds none - for a client it counts as local (loopback, a unix socket, local_addrs: the private ranges) or one that authenticated (skip_local, skip_authenticated, both on by default; milter_headers.lua skip_wanted(), lua_ip.c lua_ip_is_local()). The custom routine that signs X-SelfPost-Filter is never put through skip_wanted(), so it signed the sender's own verdict, and Inbound > Log showed it as rspamd's. Every sender is private when Docker's proxy relays port 25 from the bridge gateway. Both options set to false. Tried 2026-10-14 in a throwaway rspamd/rspamd:3.11.1-3 on the test server (rspamc --ucl, a message carrying its own verdict headers): old config, from 10.0.0.5 and as an authenticated user, the milter reply removed and added neither and added the signature; new config, both removed and rspamd's own added, as from 203.0.113.9. Also: the sender rules take the envelope only (extract_from = "smtp"; by default a bounce's From: header is used - tried: a bounce whose From: domain is denied was refused before, is not now; a denied sender is refused either way); the image pinned to 3.11.1-3, the build read (3.11 and 3.11.1 move with each rebuild); the prefilter lines no longer read go. architecture.md and security.md say so. Assumptions C-R7, B-R1, C-R4, C-R12, C-R5. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> |
||
|
mixandClaude Opus 5.5
|
57142b21b5 |
security: L8's remaining probe accepted by the owner
test / test (push) Canceled after 0s
Owner, 2026-10-14, asked whether to refuse every foreign address in a display name to close what L8 leaves (refused or passed still tells an application which domains are hosted here): "Ответ 1" - leave it. The probe is a trusted application of this server guessing names one at a time, as it already had to before L8, when the reply said in words what the line itself had named. security.md (Mail path and accepted risks, with the trigger to revisit: untrusted tenants, or the display-name rule decided anew) and the audit plan record it. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> |
||
|
mixandClaude Opus 5.5
|
ef18ef21a0 |
findings of the reviews of audit stage 2
test / test (push) Canceled after 0s
Technical review (Opus) and security review (Fable) of a59e3d4..cb5a4a2;
reports in the session's scratchpad. Nothing critical, high or medium; no
outright bug. Fixed:
- F2, F3 (0008 backfill, before it runs on the live database): rows are
filled by the rule new rows follow - the application's domain, then the
name's - and only by a domain and application that existed when the row
was written (created_at); the inbound log likewise. Test: the migration's
UPDATEs re-run on prepared rows (before the domain, after, null sender,
application gone, a stranger's name). Dry run on a copy of the live
database: 68/68 and 10/10 rows filled, none disagreeing with its login's
domain, integrity ok.
- V2, F4: an application's statistics and level-2 count are its login's
rows of its domain (appByLogin), and StatsWindow starts no earlier than
the entity's creation; the message page's "may send as" only for a login
still of the row's domain. Test: a login taken again in another domain
starts at zero.
- F7: MAIL FROM:<> is held to the domain limit of its login
(ApplicationByLogin fills DomainName).
- V3: an imported DKIM selector is held to DNS labels (validate.Selector).
- F5, F6: an imported auto limit's numbers are dropped and recalculated
right after the import; its mode is stored trimmed.
- V6: the store refuses a role that is neither global nor domain.
- F9: SaveUser keeps the stored hash when no new password is given (the
form no longer writes back the hash it read). F10: a reach naming a
domain deleted meanwhile is ErrNoSuchDomain, answered on the form.
- V4, F12: from-lines-other.txt - lines naming another sender ({other}) in
a quoted string, a quoted pair, capitals, a comment, an encoded word, a
quoted local part; the unit test wants each refused, the e2e
(lines_naming_another_sender_never_signed_for_it) wants each refused and
none signed for it. The passed list gained a.{DOMAIN}, a comment after
the angle address, a quoted local part, and "b@ {other}" - passed by
design (security.md), and the e2e shows it signed for the own domain.
- F13: e2e login_in_other_capitals_does_not_authenticate - AUTH with the
login upper-cased gets 535.
Not changed, written down: V1/F1 - L8 is half closed (refused or passed
still tells an application which domains are hosted here); security.md's
accepted risks and the CHANGELOG say so, and the owner is asked whether to
refuse every foreign address in a display name (changes the 2.4.0 rule).
F11 - deleting a domain rewrites its rows' domain_id (seconds at 1e5 rows),
accepted at this size. F8 (test cannot tell write-first from read-first;
the reviewer verified it with two stores, 30/30), V5, V7, P2: noted.
Evidence on the test server for this tree: go vet clean, go test ./... 23
packages ok, make e2e exit 0 with the three new cases.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
|
||
|
mixandClaude Opus 5.5
|
cb5a4a22e1 |
docs: stage 2 of the audit - roles, the guard, logs by id, import, logins
test / test (push) Canceled after 0s
security.md - § Roles: the last-global guard is inside the transaction that deletes or demotes (write, then count); setup is one statement; a user is saved in one transaction. A log row, outbound or inbound, belongs to a domain by id; a domain added again starts with an empty log. The Inbound › Log row of the table says by id. - § Mail path, sender check: the agreement of the check and OpenDKIM is tested (the shared list of passed From: lines); the refusal does not say which other sender a line names. - Accepted risks: "a row of the inbound log belongs to a domain by name" struck (L12). guide.md - Domain export and import: the file is checked like the forms before anything is created; a login keeps its case and a case clash is refused. - The application form: a new login is stored in lower case. schema-migrations.md was updated with 0008 in its commits. Audit plan line ticked; stage 2's lines are all done. Roadmap 22/62. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> |
||
|
mixandClaude Opus 5.5
|
c3d99b1f15 |
roadmap: decisions of 2026-10-13 on the candidates
test / test (push) Canceled after 0s
Agreed with the owner, one question at a time; documents only, no code. - csrf-tokens -> origin-strict (agreed): no tokens, the origin check refuses a request with neither header. security.md's ADR records the decision and when tokens are built after all. - structured-logging, contributing: dropped, each with what brings it back (new roadmap section "Dropped, with the condition for return"). - template-data-typing: done by 2.0.0; its last map (healthChecks) is a step of the audit's stage 4. - preflight -> health-setup-checks (certificate name, Postfix's greeting, trusted proxy as lines on Health) and test-email, both agreed. The port check is dropped with its reason; plans/preflight.md removed. - password-reset: agreed; a reset link goes only to a confirmed e-mail, so the item builds confirmation first. - phone-desktop-mode: agreed, after stages 1 and 2 of the audit. - backup-reauth (agreed): the account password before a backup or a domain export, out of review-2026-08-followups. - from-header-binding, Not done: the Delivery log box is required on a refused message's page, with the line Postfix wrote, for all three kinds of refusal (audit stage 1); the Outbound log gets a filter by status and by day (audit stage 4). The audit's checklist is 48 lines (was 45). Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> |
||
|
mixandClaude Opus 5.5
|
8f226bd82f |
release: 2.4.0: an application signs only as itself
test / test (push) Canceled after 0s
from-header-binding (docs/plans/from-header-binding.md, 9/9), at the owner's
word (2026-10-12, "Да").
- A message's From: line is held to what its envelope sender is held to:
any address of the application's domain, or the addresses of its list,
with or without a +extension. Anything else is refused with 550 5.7.1 -
another registered domain, the server's own name, a domain this server
does not sign for. Until now an application could have mail signed with
another domain's key by writing that domain into From:.
- The check is the panel's, first in the milter chain and strict: while
the panel does not answer, Postfix defers mail.
- Server > Health counts what that costs - the panel's downtime under a
running Postfix and the submissions turned away - and is laid out anew,
boxes numbered by importance.
- A refused row of the Outbound log says why in its tag.
- Migration 0006.
Upgrading: mail whose From: is not the application's own stops being
accepted - another of your domains, an address off an application's list, a
visitor's address a form puts there. The guide says what to do about each.
A stopped panel now stops submission.
Not done, written into the plan at the owner's word: the Delivery log box on
a refused message's page, the log's status filter, "messages" that are
connections, and what was reasoned about rather than run - OpenDKIM's own
parser above all.
Cut: CHANGELOG 2.4.0 written (nothing had been put under [Unreleased]
while building); image pin 2.4.0 in compose, README, guide; version table,
roadmap, plan status. security.md: a malformed packet on a milter socket,
as an accepted risk. No tag: the owner tags when he decides to, so there is
no image of this version until then.
On the test server, on
|
||
|
mixandClaude Opus 5.5
|
89fd0a6116 |
from-header-binding: findings of the technical and security reviews
test / test (push) Canceled after 0s
Both reviews were of
|
||
|
mixandClaude Opus 5.5
|
c171fb5bc7 |
from-header-binding: the sender check, what it costs, its pages; guide, architecture, security
test / test (push) Canceled after 0s
A message's From: line is held against the application that sends it, under the rule Postfix holds the envelope sender to. OpenDKIM chooses the key by that line, so until now an application could have mail signed as another domain of the server or as the server itself. The check (internal/milter/sender.go) - A role of the panel on a socket of its own (SENDER_CHECK_SOCKET), first in smtpd_milters and strict: default_action=tempfail. The journal milter keeps its socket and stays fail-open. postfix-wrapper waits for the new socket. - It reads no database: what an application may send as comes from Postfix's own sender_login_maps, what the server signs for from OpenDKIM's SigningTable. Both are read again when they change; one that cannot be read defers mail. - One From: field, one address, allowed by the envelope's lookup; then no other name of the SigningTable may appear anywhere in the raw line. Refused: 550 5.7.1 naming the line, reduced to printable ASCII. - A refusal is a send-log row with reason "from" and the header address (migration 0006), written by another goroutine and not waited for. The journal's refusals record their reason too (limit, client), and the notifier takes only a limit's. What it costs (docs/plans/from-header-binding.md, Statistics) - The role marks itself alive every 2 s; at start the gap since the mark is downtime, counted from Postfix's start if later and not at all when supervisord says Postfix is not running. - The log tailer counts the connections Postfix turned away: a warning that names the socket, then milter-reject "Service unavailable" from the same smtpd. Seen on the stand: the client is refused from CONNECT on and leaves at AUTH (454), so no MAIL line is ever written. - sender_check_days, a row per UTC day, kept a year. Pages - Health: the card "Sender check", the socket's row, and the cards layout - boxes numbered by importance in two stacks read in rows; the left stack polls, the three readings of the right one are swapped in by id. - Outbound log: a refused row says why in its tag (wrong From, wrong client, rate-limited); the message's page shows the From: line and what the application may send as. - Help: topic "Sender check", word for word the mockup's. Found by the stand, fixed here: the names for the second look came from the sender bindings, and a domain with a key and no application yet was not among them. e2e: negative_journal_milter_fail_open is gone - it proved that mail passes with the panel down, which is no longer so. In its place sender_check_holds_the_from_line (another domain's From:, the hostname's, a stranger's, an address off the list, another sender in the display name - refused; own addresses - delivered) and panel_down_defers_mail (4xx with the panel stopped, the container alive, mail again after, downtime and one deferred message on Health, the card warning). Test server, this tree before the last cosmetic edits (a fixture row, a fact's size, docs): go vet ok, go test 23 packages ok, make e2e ok (337 s). Locally after them: internal/web/view ok. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> |
||
|
mixandClaude Opus 5.5
|
f09b60aaa2 |
panel-notifications: findings of the technical and security reviews
test / test (push) Canceled after 0s
Neither review (read-only agents, Opus and Fable, over 972a9f8..400aa76) has a Critical finding. Both found the same serious one, and more: - The notification about refused mail was built on what a rejected row says of itself. Its domain is the one in MAIL FROM, which the client typed and Postfix has not yet held against the login when the milter refuses: an application over its limit could have had the panel write a message per made-up domain, to another tenant's users among others. And the same status is written when an application is refused for its client address, which is no limit. Refusals are now counted by the domain the login is registered under, only for logins that exist and only while a limit is set on the login or its domain; one subject per sending domain. They are read through a partial index of their own (migration 0005): the query walked the whole send log every minute. - Account and user e-mails were checked with the DMARC report validator, which refuses Gmail and the like - the owner's own kind of address. A new validate.AccountEmail accepts them and refuses a local part beginning "-". - "Said" now means sent: what arises while nobody can be told, or while Postfix does not take the message, is told at a later check. Its ending with nobody left to tell is simply over. - Two DNS checks must be 45 minutes apart as well as two: the last check before a restart and the first after it were minutes apart. - The notifier's own mail waiting in the queue, and its bounce coming back to postmaster@ this server, are not the queue's subject. A listing cut at its size limit or with a line too long to read is no listing; postqueue -j gets 30 seconds. - The tables must name the selector the panel shows, not only have the line. - An a/mx lookup that did not answer inside the SPF check made the record "not authorising"; it is "cannot tell" now, for the pages too. - A ceiling of 60 messages an hour; a pass stops between domains when the process is asked to stop; sendmail gets a WaitDelay; days are rounded as a person counts; a failed read of the switch shows it on, not off. Recorded as accepted in security.md, with a roadmap candidate (from-header-binding): an application can have a message signed as the server's own name - or, as always, as another registered domain - by writing it into the From: header, which nothing holds against the login. Left as found: a recipient whose copy failed while another's went out is not tried again; a certificate that lives under 14 days is reported at each renewal. On the test server, for this tree: go vet clean, go test ./... 23 packages ok, make e2e ok (326 s). Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> |
||
|
mixandClaude Opus 5.5
|
c9027381e5 |
panel-notifications: guide, architecture, security; refusals looked for every minute
guide.md gains Notifications: the four things, when each is said and said again, who is told, the two switches, and the two DNS records of the server's own name that make the mail arrive. architecture.md: the notifier as the fifth role, its readings, state, recipients and sending; the server's own key in OpenDKIM's tables; non_smtpd_milters. security.md: the panel submits mail of its own - how, to whom, and what text from outside can reach a body. That text - a queued message's sender, a DNS finding - is reduced to printable ASCII and cut before it is printed. Refused mail is looked for every minute instead of every five: it is one indexed query, and a refusal is mail not being sent now. The image's note on the panel's postdrop membership no longer calls it read-side only. go test ./internal/notify passes. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> |
||
|
mixandClaude Opus 5.5
|
a990e7cc23 |
delivery-log-storage: findings of the technical and security reviews
test / test (push) Canceled after 0s
Neither review (read-only agents, Opus and Fable, over ae06d8b..815236a) has a Critical or High finding. What they found and what was done: - A stranger on port 25 could grow the table: every queue-id line of inbound mail was stored and pruned an hour or more later, his own message-id among them. The tailer now leaves out the lines of a queue id that postfix/inbound/smtpd opened, until qmgr removes the message or another listener opens one under the id; 4096 such ids are remembered. - A commit, so a disk sync, per stored line on the connection the journal milter shares, and a backlog written line by line. Lines are now written in one transaction per pass over the log, 500 at most, before the read offset is saved (InsertSendLogLines; follow() calls the keeper's flush). - The prune asked the send log once per line and ran hourly. It asks once per queue id and runs at start and with the six-hourly sweep; with port-25 mail left out there is little without a row to wait for. - The 4096-byte cut could split a character; it keeps characters whole. - security.md did not say that a backup now carries these lines, and the plan said the time window excludes a reused queue id where it only limits the chance to the row's queue lifetime. Both are corrected; the collision is recorded as an accepted risk with its trigger. - Tests that did not prove their comments: the prune is now tested with the sweep, a rejected row next to the prune, batching and the change from catching up to following, port-25 mail left out and its id reused. Left as found and written down in the plan: lines the tailer never reads (a file rotated away while the panel was down), and a message in the queue across the upgrade, which shows only its later lines. On the test server, for this tree: go vet clean, go test ./... 22 packages ok, make e2e ok. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> |