swag v2.0.0-rc5's parseSecAttributesV3 resolves a security scheme's map key
via getSecurityDefinitionKey(lines), which scans from the start of whatever
comment-line slice it was handed and returns the first @securitydefinitions
match — ignoring the current parse position entirely. Three
@securityDefinitions.apikey blocks stacked in one Go comment group (the
three were separated only by bare '//' lines, which do not split an
ast.CommentGroup) therefore all resolved to the first block's name
(cookieAuth), with the last block's in/name/description winning: the
generated document had exactly one securityScheme, keyed cookieAuth, body
esoAuth.
Separating the three blocks with real blank source lines splits them into
three distinct ast.CommentGroups, so swag's file-level comment scan (which
requires no other tokens between them, same rule Go uses for doc comments)
hands each block its own line slice and each resolves its own key.
Regenerated openapi.json now carries all three schemes with correct
bodies, referenced with no dangling security requirements.
Generated from swaggo v2 annotations, committed rather than built into the
image: the runtime stage is scratch and adding codegen puts the toolchain
in the build. CI regenerates and diffs, so an annotation edited without
regenerating fails the build — without that the annotations would drift
while still looking authoritative.
Scalar is vendored rather than loaded from a CDN, because air-gapped
self-hosted installs are supported and a reference page that fails closed
offline is a support ticket.
Same treatment: named types replace gin.H literals, and every handler gets
a swaggo doc block. This is the last of the handler files under
server/internal/api/.
Same treatment: named types replace gin.H literals, and every handler gets
a swaggo doc block. VulnSummaryResponse uses pointer fields so the
db-freshness block stays entirely absent when no vulndb_meta document
exists yet, matching the handler's original conditional gin.H exactly.
Converts their gin.H responses to the named types added in the previous
commit and adds swaggo doc blocks for every handler in handlers.go and
tokens.go.
Named response types for handlers that were returning anonymous gin.H
literals, so a generated annotation and what the handler actually returns
cannot disagree. main.go carries the top-level swaggo info block (title,
description, security schemes for cookie, bearer token and ESO auth).
A card in the Access group beside Members and single sign-on rather than a
new nav entry — /settings/instance was folded back in for exactly this
reason. The plaintext is shown once in a well block and never again.
Tokens outside a newly tightened lifetime policy are flagged rather than
broken, because the policy governs issuance, not existing credentials.
600 per minute per token, in the Redis that sessions already require.
Cookie sessions are untouched. A Redis failure falls through rather than
refusing traffic — it is already a larger problem and should not become a
second outage.
createToken's catch-all mapped every unmatched error to 400, so a
database outage reported itself as a malformed client request. Wrap the
genuine validation failures with ErrTokenInvalid and let the handler
answer 500 with a fixed message for everything else.
createToken's catch-all answered 400 for every unmatched error, so a
database failure reported itself as the caller's malformed request. Found
in review of Task 7.
Create, list and revoke, with no update: editing what a credential already
deployed in CI can do, with no record of what it could do before, is worse
than requiring a rotation. Revoking a token that is not yours answers
not-found, since a 403 confirms it exists.
The audit actor stays the human and names the credential alongside, so a
person clicking and their CI job are told apart.
Keyed on the registered gin route pattern rather than a per-route
decorator, because a route registered without a decorator would be
unguarded. An unmapped route reached by a token is a 403, and a boot-time
check refuses to start when any /api route is missing, so the failure
lands at deploy rather than as a customer's surprise 403.
sessionFromCookie already writes "session expired" when a cookie was
presented and rejected with no bearer to fall through to. Middleware
called sessionFromToken anyway, which wrote a second "not authenticated"
body onto the same response for every ordinary browser-session timeout -
gin logged "superfluous response.WriteHeader call" on ordinary use, not a
rare edge case.
Guard on c.IsAborted() after sessionFromCookie: true only in that one
rejected-cookie-no-bearer branch, so it short-circuits there while the
other three credential paths (no credential, bearer only, stale cookie
plus valid bearer) are unaffected.
The plan's Middleware called sessionFromToken even when sessionFromCookie
had already answered a rejected cookie, putting two JSON bodies on the
wire for the ordinary expired-session case. Found in review of Task 5.
One middleware, two ways to arrive at the same *Session, so every handler,
role guard, licence gate and audit call is untouched. The host guard
applies to both: a token carries an instance, and the tenant boundary must
not have a token-shaped hole in it.
The effective role is min(user, token) recomputed per request, so demoting
somebody demotes their tokens with them. A stale cookie beside a valid
bearer falls through rather than refusing a credential that would work.
Mint, resolve, list and revoke, with the effective role capped at the
owner's and recomputed per request rather than frozen at creation.
Deleting a user deletes their tokens in the same call, so offboarding is
one action. Revoking somebody else's token answers not-found rather than
forbidden, since a 403 confirms the credential exists.
Also re-exports shared.APITokenMaxDays into server/internal/models,
following the existing ValidRole wrapper pattern, since the token
service needs it and it was never re-exported.
A pointer with absent meaning no cap, so an upgrade allows never-expire
tokens exactly as before and an instance opts into the policy. It governs
issuance only: changing it never invalidates a token that already exists.
Eight resources with read and write, write implying read. Coarse on
purpose: a scope per endpoint is a table nobody maintains, and a route
added without an entry either breaks or is unguarded.
The unique index on token_hash is what makes authentication an indexed
lookup rather than a scan, so this builder is fatal on failure like
EnsureAuthIndexes rather than warning like the secrets one.
Registered in ScopedCollections so instance purge reaches it.
Twelve tasks from model through middleware, scope enforcement, endpoints,
web UI, generated OpenAPI and documentation. Verification is build plus
curl and UI checks rather than test cycles, matching a repository with no
Go test harness beyond shared/mail.
Adds the approved design for personal access tokens on the control plane
REST API, and for the generated OpenAPI 3.1 document served as a Scalar
reference page.
Tokens fall back into the existing session middleware rather than getting
their own route group, so every handler, role guard and audit call works
unchanged. Scope enforcement derives from the registered route pattern and
fails closed, with a boot-time check for unmapped routes.
A Terraform provider is deliberately left to a follow-on spec.
The page rendered a map of eleven event types to labels and seven to colours.
The server emits forty-seven. Everything unmapped fell through to its raw
string, so "Key Assigned" in green sat above "workflow.schedule_updated" in
grey — the same kind of fact in two formats, which made the column look like it
carried a meaning it did not.
Presentation is now derived rather than enumerated. Event types are named
<category>.<action> by every call site, so the category becomes a chip, the
action is humanised, and the tone comes from the verb. A type added to the
server tomorrow gets a sensible label and colour with no second list to update;
the override table holds only the dozen the rule reads badly for. Every row is
one treatment, and colour never carries meaning alone — the sentence beside it
says the same thing in words.
Paging and filtering are server-side, unlike the fleet lists that answer with
everything and slice in the browser. audit_retention_days is a licensed
entitlement measured in months, and this log is read to answer questions about
the past, so a browser filtering the most recent page would report "no results"
for events that exist. GET /api/audit now takes q, category, limit and skip and
answers {events, total} — a short page is not evidence of the end of the log,
which is why the total is counted rather than inferred.
audit_logs had no indexes at all: every read was a collection scan with an
in-memory sort over an append-only collection. Adds (instance_id, created_at)
and warns rather than failing, matching EnsureSecretIndexes.
Two bugs found by running the deriver over all forty-seven real types rather
than eyeballing it: the tone rules matched only past-tense verbs, leaving
auth_provider.delete drawn as neutral beside key.deleted in red; and
"unaccepted" matched "accepted", so withdrawing an acceptance read as the same
caution as granting one.
The last of the inline two-step deletes, and the one with the most reach: a
key delete revokes it from every server at once, and for a generated key the
stored private half goes with it. That copy is the only one Vantage holds, so
unlike an uploaded key this cannot be undone by pasting the public half back.
Body names how many servers lose access, and says the private key is destroyed
only when there is one to destroy.
The delete error moves out of the toast and into the dialog, which stays open
on failure, matching the server and monitor deletes.
Destructive actions were confirmed by a second danger button rendered where
the first one had been, so a double click on Remove deleted the thing without
the operator ever reading which thing it was. Servers, monitors, secret keys
and sign-in providers now go through ConfirmDialog with requireTyped, matching
the secret-group delete that already worked this way. Notification channels and
vulnerability alert rules get an untyped dialog: both are a name and a URL and
are rebuilt in a minute, but neither had any confirmation at all, and both
silently stop alerts that nobody misses until an incident goes unannounced.
Mutations otherwise succeeded in silence, or reported into whatever inline
banner the page happened to own. Two failure modes came of that: a modal that
closed on error left the message nowhere to land, and a save that was rejected
left the old values on screen looking exactly like a save that worked
(/settings had no error path at all). Every mutation now reports through the
existing toast context.
Errors stay inline where the surface that raised them is still on screen and
the message is a correction to make in it: form validation, the cron field,
the tag rows, a rejected licence blob, and the workflow designer's autosave,
which is a standing condition rather than an event. Everything else toasts.
Ad-hoc feedback removed in favour of it: the "Sent!" button labels on the
server maintenance tab, the settings "Saved!" flag, the channel test line, and
the steps page's notice/error pair.
Both from the final review, and the first one overturns a call I got wrong.
The aria-hidden sweep that makes aria-modal true also swallowed the toasts.
ToastProvider renders inside the app root, and every modal-raised confirmation
is toasted *before* its dialog closes — "Saved …", "Deleted …", "Removed …" —
so each one was inserted into a hidden subtree and never announced. Un-hiding
a live region afterwards does not replay what it missed. The toast layer is
portalled to the body carrying the dialog-layer attribute, which exempts it
from the sweep, and sits above the dialog: a toast explaining why a dialog's
action failed is no use behind it.
EmptyState's action was narrower than the call site it replaced. The old
first-workflow button carried loading={isPending}; the new one carried
nothing, so a double click created two workflows. The action is a union now —
a link takes no pending state, a handler takes loading and disabled.
Keys, secrets, audit, steps and workflows each carried the same
loading/error/empty ternary with its own copy of the spinner and its own
wording for a failed fetch — five of them had drifted to five different
sentences for "the request did not come back". They go through AsyncBoundary
now, which means a skeleton in place of a spinner, a retry button on failure,
and backend messages passed through friendlyMessage rather than printed raw.
Steps gets the filtered-empty state the fleet just got: "no steps match that
filter" is not "no steps yet", and only one of them should offer to create the
first one.
Two more from review:
Modal's effect ran before `mounted` flipped, so a dialog rendered already open
found null refs and took no focus at all. It depends on `mounted` now.
aria-modal was a claim with no mechanism behind it — portalled to the body,
the app tree is a sibling of the dialog and a screen reader's virtual cursor
still browsed the page underneath. The body's other children are marked
aria-hidden while any dialog is open, refcounted alongside the scroll lock.
This does mean a toast raised while a dialog is open is not announced, which
is the correct trade for a modal: ConfirmDialog shows its own errors inline.
Two of these were real defects in the previous two commits.
friendlyMessage discarded exactly the messages it claimed to keep: the
"is this a bare reason phrase" test was a shape regex, and "Default steps
cannot be edited" has the same shape as "Not Found". It is an exact-match set
of reason phrases now.
Modal depended on onKeyDown, which is rebuilt whenever onClose changes
identity — and onClose is an inline arrow at every call site, so any parent
re-render (a 30s poll, a mutation flipping to pending) tore the effect down
and rebuilt it: cleanup restored focus to the trigger, setup then moved it to
the top of the dialog, mid-typing. onClose is held in a ref and the effect is
keyed on `open` alone.
Also in Modal: initial focus takes the first control in the body rather than
the panel, since the header comes first in DOM order and every dialog was
opening on its own dismiss button; the Tab trap pulls focus back when it has
escaped the panel entirely rather than only handling the two ends; the scroll
lock is refcounted, because per-instance save/restore released the page when
an outer dialog unmounted under an open inner one; and the whole thing is
portalled to the body so a nested confirm is not clipped by its parent's
overflow box.
The fleet's filtered-empty state keyed on the search alone, so a tag filter
matching nothing told a customer with a full fleet to add their first server.
Remaining: pending mutation errors are reset when a confirm dialog closes, so
one member's failure no longer greets the next; ConfirmDialog clears typed
confirmation when the target changes, not only when it reopens; deleting a
workflow closes its dialogs before navigating rather than carrying a scroll
lock onto the next page; toasts split into a polite and an assertive region,
since one polite wrapper demotes the role="alert" children inside it; and a
custom skeleton gets a live "Loading" beside it, having been aria-hidden with
nothing else to announce.
The fleet list had a tag filter and nothing else: no search, no sort, and an
unbounded list. Searching hostname/address/OS and sorting by hostname, status
or last seen are all client-side, since the browser already holds the fleet
the page just fetched. Sorting by status orders by how much attention each
state wants rather than alphabetically, which is the only reason to sort by it.
The filtered count is shown beside the total so a search does not read as the
fleet having shrunk, and "no results" is a distinct empty state from "no
servers", with a way back out of the search.
refetchIntervalInBackground defaults to false on the query client. Polling
pages kept refetching in a hidden tab — the fleet list pulls inventory blobs
every 30s — so a console left open in a background tab polled until its
session expired. It belongs in the defaults because the argument is identical
on every polling page.
Four correctness/accessibility defects and the destructive-action flow.
- Button: the loading spinner carried xmlns="http://www.w3.instance/2000/svg",
a find/replace of "org" that landed inside a URL. Button also grows an href
form, because <Link><Button> nested a button inside an anchor at nineteen
call sites: invalid markup, two tab stops, and Enter firing only the anchor.
- Fleet status was four meanings carried by hue with the distinction living in
a title attribute, which touch never shows and screen readers need not
announce. It now carries a text label and an accessible name, which is the
one rule the design system states outright.
- Modal had no focus management at all: no trap, no initial focus, no restore,
no scroll lock, no aria-labelledby. Dialogs nest (a confirm over an edit), so
a stack decides which panel owns Escape and Tab.
- Seven destructive actions went through window.confirm(). ConfirmDialog
replaces them and can say what is about to happen; deleting a secret group,
a shared base step or a workflow now requires typing the name, since those
have no undo and a wide blast radius. adminsite keeps its own inline idiom
rather than importing a dialog system it does not have.
Adds Toast, AsyncBoundary/EmptyState/ErrorState/TableSkeleton and
friendlyMessage, replacing per-page loading ternaries and raw
(error as Error).message text. Wired here only where a call site was already
being edited; the remaining pages follow.
Every other page under app/(app) opens with `p-4 sm:p-6 lg:p-8` and the
layout adds none of its own, so this page alone sat flush against the shell
edge. Its h1 was text-xl where every other page is text-2xl.
The findings list was a stack of separately bordered cards; it is now rows
inside one Card, separated by border-border-soft, matching the monitors and
workflows lists. Loading is the shared spinner rather than a line of text,
the error is the shared danger strip, and the empty state uses the same
proportions as the monitors one.
The runtime stage is FROM scratch, which has no /tmp, so vulnsched died at
startup with "temp dir: stat /tmp: no such file or directory" and no scan
ever ran. Nothing in the server wrote to a temporary directory before the
trivy-db puller, which is why this only appeared now.
scratch cannot mkdir its own, so the directory is staged in the builder at
1777 and copied in. Also corrects CLAUDE.md, which described this image as
Alpine; the time/tzdata import it justifies is if anything more load-bearing
on scratch.
trivy-db, oras-go, the three version comparators and robfig/cron were
recorded as indirect with an incomplete go.sum, which builds locally
against a warm module cache but fails in CI with "updates to go.mod
needed". trivy-db pulls testify into the build graph, and its hashes
were missing entirely.
The catalogue row alone was not enough; the feature was invisible in
three places and mislabelled in a fourth.
PlanConfigurator rendered every key that was not "console" as "Single
sign-on", so the staff checkbox granting vulnerability scanning was
labelled single sign-on. Feature wording was duplicated between the staff
configurator and the purchase form and the copies had drifted, so it now
lives in adminsite/lib/features.ts and both read from it.
The customer licence panel showed raw keys; it now labels them.
Pricing gains a comparison row. The add-on block with a monthly price is
deliberately NOT added: that is a pricing decision, and the Paddle price
IDs for the new catalogue rows have to be pasted in before it can be sold
anyway.
Adds license.FeatureVulnScanning as the one name for the feature and a
catalogue row per deployment/tier, following console and oidc: features
are opt-in per customer, so no plan bundles it.
Documents the subsystem in CLAUDE.md, including that ScopedCollections is
the canonical registry instance deletion derives from — there is no
separate deletion list, which the plan had wrong.
Fleet board grouped by CVE, a per-server section on server detail, and
alert rules beside the channels they consume.
The server detail page has no tab pattern despite the plan saying to
follow one, so this adds a section in the existing vertical stack.
Three states are kept visually distinct because they are identical if
handled carelessly and only one is good news: never reported, no advisory
feed for the distribution, and scanned-and-clean. Database freshness sits
with the findings rather than in settings for the same reason.
Completes tasks 10-15 and fixes what was outstanding:
- vulndb.Pull implemented with oras-go, streaming the ~50MB layer and
staging both files before replacing either, so a failed pull leaves the
previous database intact rather than a half-written one.
- db.go: Vulnerability.Severity is a string, not trivy Severity, so the
int conversion did not compile. Severity now resolves vendor (highest
when vendors disagree) then NVD then unknown, and CVSS is read too.
- findings.go: added sweepFixedFindings plus the fleet query, severity
counts, rescan flag and accept/unaccept the API needs.
- vulnrules.go: added rule CRUD and the digest builder. ResolveTargets
returns []models.Server, not []string, so filterByServers was wrong.
- api/vulnerabilities.go was an empty file while handlers.go registered
twelve routes against it; written, grouped by CVE.
- shared/mail: added the missing sender. The templates were orphaned and
the HTML one was a copy of the text one, defining "subject" (which
html/template would escape) and emitting no markup. render.go parses
every template in init(), so a bad one panics server, admin and sitesvc
at boot — go build never runs init(), which is why nothing complained.
- notify: digests dispatch through their own path so SMTP gets the digest
template rather than arriving dressed as a monitor alert.
SyncKeys now returns the whole response so the poll can carry
CollectPackages; a separate RPC for one boolean would be a message every
30 seconds for a value that changes when a licence does.
The flag is an atomic: the 30s poll writes it, the hourly package loop
reads it, and they are different goroutines.
VulnScanningEnabled reads GetLicenseState(...).Feature("vuln_scanning")
and requires an active licence, never switching on tier. ReportPackages
re-checks it server-side: the agent flag is the optimisation, this is
the boundary.
The pb packages are hand-written, not protoc-generated, and the wire
codec is JSON (encoding.RegisterCodec(JSONCodec{})). Field numbers in
the .proto are documentation; JSON field names are the contract. Both pb
packages edited by hand to match.
SyncResponse.collect_packages is omitempty and absent decodes as false,
so an older server leaves agents collecting nothing rather than
collecting without a licence.
Adds server_packages, vuln_findings and vuln_alert_rules to
ScopedCollections rather than to a separate deletion list. purgeInstance
derives its collection list from that registry, so instance deletion
follows automatically and there is no second copy to drift.
Both plans now verify by build, vet and manual checks written into the
tasks. Spec verification sections updated to match so they no longer
describe tests that will not be written.
Agents enumerate Docker containers, compose stacks and systemd services;
start/stop/restart and bounded log snapshots from the UI.
Sub-project B, Linux only. Live log following stays in the console.
17 tasks, TDD where the logic is pure. Corrects two spec claims:
the server reads features via License.HasFeature rather than admin's
entitlement directly, and shared/mail/render_test.go does not exist.
Agents report installed packages; the control plane matches them against
trivy-db and raises findings that link to the existing ApplyUpdatesCmd
patching path.
Scoped to sub-project A, Linux only. Container registry, image scanning
and compliance baselines are separate specs.
mrhid6