Extends IZaakSource with RecordBesluit, mirroring WP-50's CreateZaak write
pattern: OpenZaakZaakSource POSTs a new Statussen entry (highest-volgnummer
statustype, since the harness catalogus has no per-outcome besluittype),
carrying the besluit + toelichting in statustoelichting; LocalZaakSource
no-ops. The beoordeling endpoint calls it after the local decision commits,
flagging a failure via RecordZgwDivergence the same way submit's
create-zaak/document writes do — closing WP-60's "second write pair" gap.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
WP-67's monorepo split renamed .storybook to .storybook-ssp/
.storybook-behandelportal, but test-storybook still defaulted to the
(now nonexistent) plain .storybook dir -- it loads <config-dir>/main.js
for test-runner hooks even in --url mode, so every invocation failed
with "Could not load main.js in .storybook".
Also: test-storybook:ci only ever built+served+tested the ssp instance.
Since the split, behandelportal's stories (werkvoorraad, beoordeling,
besluit-form, ...) were never axe-tested in CI at all. Added the
:behandelportal siblings (mirroring the existing storybook/
build-storybook naming) and wired them into ci-local.sh's --full step
and the storybook-a11y GitHub Actions job.
Verified directly: ssp 62/180 stories green, behandelportal 45/112
green (including the new besluit-form story), full `ci-local.sh --full`
green end-to-end.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
WP-65's summary-table row still said "in progress (65a done)" from before
65b shipped, and that one oversized cell was forcing prettier to want to
re-pad the entire ~65-row table. Correcting it to "done" (its actual
status, per the WP-65 file's own outcome notes) removes the outlier —
npm run format:check is green again with zero other rows touched.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Adds POST /beoordeling/{id}/besluit: a Besluit enum (Goedkeuren/Afwijzen/
MeerInfoOpvragen) backed by new Aanvraag.BesluitStatus/BesluitToelichting
columns, gated by the same BeoordelingRules.CanDecide the read side's
canBesluiten flag already uses (409 on an illegal transition, 400 on a
missing required toelichting). Mappers.ToStatusDto gains the "a recorded
decision wins" branch. FE: besluit.machine.ts + besluit-form organism
(same form idiom as change-request-form), wired into the beoordeling page
behind the server's canBesluiten flag.
Completes WP-65 (65a + 65b) — verified end-to-end against a running
backend (werkvoorraad -> beoordeling -> besluit -> status reflected back).
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
New GET /beoordeling/{id} shows one aanvraag's status, linked documents, and a
canBesluiten decision flag, gated by the same CanBeoordelen capability as the
werkvoorraad list. Reads through IZaakSource.ListCases rather than a new seam
method (WP-66 needs one anyway for the real write); owner BSN is masked.
Fixes a real gap found while wiring this up: the behandelportal's login was still
WP-61's copied citizen/BSN DigiD flow, so nothing ever sent X-Medewerker and the
werkvoorraad screen (WP-64) always denied in a real browser. A dev-only
medewerkerInterceptor (mirrors the existing ?role= stand-in as ?rollen=) fixes that.
WP-65's own Risks note authorized splitting read from write across sessions given
its size; this is the read half. The decision-recording mutation is next (65b).
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
New GET /werkvoorraad endpoint lists aanvragen still open (Ingediend/InBehandeling),
gated by the medewerker capability (CanBeoordelen) rather than the admin role — reuses
the existing ApplicationSummaryDto, no new DTO. GET /me now surfaces aanvraag:beoordelen
for a behandelaar so the FE can gate with the same AccessStore/capabilityGuard idiom
every other page uses.
FE: a behandeling domain type deliberately narrower than ssp's full AanvraagStatus
union (only the two open tags — illegal states unrepresentable), composed into a
werkvoorraad-list organism from existing shared/ui molecules. Replaces WP-61's
scaffold placeholder as the app's real landing page.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Restructures into apps/ssp + apps/behandelportal (two Angular projects)
plus libs/shared + libs/beheer (cross-app libraries), replacing WP-61's
separate sibling repo. That split had already produced real drift: a
hand-vendored copy of the backend's OpenAPI doc, a shared/ui+layout tree
forked and silently diverging (7 files), and beheer + the styles.scss
token bridge duplicated byte-for-byte across both repos.
- git mv the SSP's src/app/* into apps/ssp/; fold shared/, beheer/,
environments/, the Storybook docs/*.mdx, and styles.scss into
libs/shared + libs/beheer (all confirmed identical between the two
repos before merging). auth stays deliberately duplicated per
ADR-0002 (actor-specific, expected to diverge) - amended there.
- One generated API client (libs/shared), no more vendored swagger.json.
- .dependency-cruiser split into a base factory + one config per app,
and Storybook into .storybook-ssp/.storybook-behandelportal - both
forced by the @auth/* alias resolving to different directories per app.
- SiteHeaderComponent/ShellComponent gained HEADER_NAV_ITEMS/
HEADER_ADMIN_LINKS/DEBUG_PANEL injection tokens so each app supplies
its own nav/admin-links/dev-panel instead of one being hardcoded.
- CLAUDE.md, ARCHITECTURE.md, dependencies.md, and ADR-0002 updated;
WP-67 backlog entry documents the full decision trail.
npm run ci green (lint, dep:check x2, 360 tests across ssp/
behandelportal/shared/beheer, both localized builds, backend tests,
snippet + api-client drift); both dev servers, both Storybook
instances, and docker compose verified working.
The old sibling repo (/home/eho/repos/behandelportal) is left
untouched, not deleted.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Model the full ADR-0002 lifecycle (Ingediend/InBehandeling/
MeerInfoGevraagd/Goedgekeurd/Afgewezen) as a backend enum backing the
existing AanvraagStatusDto.Tag string, and widen the FE union/parse
boundary/switches to match. Ingediend/MeerInfoGevraagd aren't reachable
yet (no behandelaar transition exists) — that's WP-65. Zero DTO shape
change, so gen:api has no drift.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Splits backend CallerIdentity into the two ADR-0002 §3 actor kinds
(ZorgverlenerCaller/MedewerkerCaller), a stub X-Medewerker/X-Rollen header
path mirroring WP-53's citizen stub, and Authz.CanBeoordelen as the first
medewerker capability — backend-only, no consumer until WP-64. Also fixes
the backlog README's stale WP-61 status (done, but table said todo).
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
WP-61 (bootstrap the behandelportal app) is done — a separate sibling repo
at /home/eho/repos/behandelportal, not a commit in this one. Records the
one real kickoff deviation from the vanilla new-ssp recipe: create-ssp.mjs
unconditionally renames the backend project even with --skip-backend, which
conflicts with the WP's "no new backend service" decision. Resolved by
deleting backend/ from the new repo and vendoring its swagger.json into
api-contract/ instead, with gen:api regenerating only the client against
that vendored doc.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Renamed scripts/create-ssp.mjs -> create-frontend.mjs (+ its WP-45 doc, npm
script, and every prose/command reference) since "ssp" reads as an acronym
where "create-frontend" says what it does.
Also fixes two real bugs found while running it for real during WP-61:
scripts/ci-local.sh was missing from RENAME_CONTENT_FILES (any --name'd
clone that keeps a backend would break `npm run ci`, still hardcoding
BigRegister.slnx), and plopfile.mjs's `gen:context` insertion into
.dependency-cruiser.js anchored on the `showcase: null,` line, which
create-ssp/create-frontend has already stripped by the time gen:context
runs in the same invocation — silently leaving a freshly scaffolded
context with no CONTEXT_ALLOWED fence entry at all. Re-anchored on the
`const CONTEXT_ALLOWED = {` line instead, which never moves.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
scripts/openzaak-ui-up.sh brings up the root app and the OpenZaak harness
together, wires them onto one docker network, seeds the catalogus, grants
the container-alias zaaktype scope, and verifies a real aanvraag submitted
through the UI lands in OpenZaak.
Along the way: DocumentStore.DemoOwner was reusing the seeded doctor's
11-digit BIG-nummer as a stand-in BSN, which isn't a valid 9-digit BSN shape
— OpenZaak rejects it, breaking both submit's rol-creation step and the
citizen's own applications list under Zgw:Enabled=true. Fixed to a real
elfproef-valid BSN.
Also adds mitigation for a still-unexplained per-container flake (every
outbound ZGW POST fails as if the body were empty, for that container's
whole lifetime) that correlates with host memory pressure: the script now
warns when host swap is heavily used, and an opt-in ZgwDiagnosticHandler
(ZGW_DEBUG_HTTP=1) logs Content-Length vs. actual bytes sent so the next
reproduction can confirm or rule out client-side body corruption.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Local aanvraag/document writes and their paired ZGW writes aren't
transactional; a ZGW failure after the local write succeeds used to
diverge silently. ZgwHttpClient now retries transport-shaped failures
(not 500, which can follow a partial commit on the non-idempotent
statussen/rollen POSTs), and a ZGW failure that survives retry sets
Aanvraag.ZgwError plus a zgw:divergence audit row instead of failing
or diverging quietly. No outbox/reconcile job: three request-triggered
write paths don't justify a persisted queue that would also need to
carry citizen PII for the JWT audit claims.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Drives the DRC upload's vertrouwelijkheidaanduiding from a new stamdata
table instead of the hardcoded "openbaar", following the existing
config-as-code pattern (ADR-0004). Adds the referential-integrity check
StamdataValidationTests was missing for the new table.
OpenZaak doesn't serve the Notificaties API itself (it's a separate app,
open-notificaties) — standing one up for a real abonnement would triple
this harness for a benefit it doesn't need (exactly one subscriber, this
repo's own BFF). Instead, an opt-in compose overlay adds a celery worker
and points OpenZaak's NotificationsConfig straight at the BFF's webhook
via a zgw_consumers Service; bootstrap-notificaties.sh configures it
idempotently and verify-notificatie.sh proves a real write delivers to
the BFF's audit trail end-to-end.
Verified live: preflight proves the webhook's shared-secret gate both
ways (204/401), a zaak PATCH triggers real celery delivery, and rerunning
both scripts against an already-configured harness stays idempotent.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
setup_configuration has no YAML field for granular autorisaties, so
bigregister-test now starts at heeft_alle_autorisaties: false (dev + prod
template) and bootstrap-catalogus.sh grants exactly the ztc/zrc scopes the
harness needs via the Django ORM, sidestepping the zero-scope
chicken-and-egg with the JWT-authenticated Autorisaties REST API.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
bootstrap-catalogus.sh now looks up every resource by its natural key before
creating it (catalogus by domein+rsin, zaaktype by catalogus+identificatie,
statustype by zaaktype+volgnummer, roltype by zaaktype+omschrijvingGeneriek,
zaaktype-publish by checking `concept` first, zaak by identificatie,
status/rol by existence-under-the-zaak), so rerunning against an
already-seeded instance reuses what's there instead of erroring.
The WP's original plan (move this into OpenZaak's `setup_configuration`
mechanism) turned out not to be achievable: reading the actual
django_setup_configuration steps installed inside the open-zaak image shows
no step exists for Catalogi/Zaken content anywhere in this OpenZaak version
— only sites/credentials/applicaties/selectielijst. Documented as a
deviation; the WP's own Risks section already anticipated this and sanctioned
falling back to an idempotent script.
Verified live: fresh instance -> full run (all created) -> integration test
green -> reran the script twice more against the same instance (all reused,
identical URLs, no duplicates) -> integration test still green.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
The rijks-sans-regular/italic.woff2 files were never vendored (only
CIBG-icons-1.2.woff2 exists), so the @font-face rules 404'd on every page
load and the font stacks named a font that could never load anyway. Matches
the already-documented decision to run a system-font stack (licensed RO/Rijks
fonts not shipped).
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
docker-compose.openzaak.prod.yml layers real SECRET_KEY/DB password/site
domain/allowed-hosts (all required, fail-fast via ${VAR:?...}) on top of the
WP-54 dev harness, switches Postgres off trust auth, and sets IS_HTTPS for a
front-facing reverse-proxy TLS setup. The ZGW client secret lives inside a
file setup_configuration reads rather than a compose env var, so it's
templated (data.prod.yaml.template, no secret) and rendered host-side via
render-prod-secrets.sh into a gitignored data.prod.yaml, mounted over the
container's dev data.yaml. ZgwOptions.cs already binds from IConfiguration,
so the BFF side needed no code change.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
WP-55..60 harden the OpenZaak integration for production (secrets/TLS,
idempotent provisioning, least-privilege scopes, real notifications,
confidentialiteit config, write-divergence resilience). WP-61..66 stand up
a staff-facing behandelportal per ADR-0002, wired to the same backend via
BFF-lite decision DTOs. Both phases are independent tracks; WP-60's
Decisions block is deliberately left open for a planner-agent kickoff.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
The watched push came back all-green after ebf1f8f (the non-root Dockerfile
fix semgrep's live run caught). Marks the entire showcase/ZGW backlog
(WP-05..54) done.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
The pushed WP-30 item-5 Dockerfile predated the semgrep triage's local run —
CI's now-blocking semgrep gate caught what local verification couldn't:
dockerfile.security.missing-user-entrypoint (no USER, container runs as root).
mcr.microsoft.com/dotnet/aspnet:10.0 ships a pre-created non-root user for
exactly this ($APP_UID, uid/gid 1654) — switched to it, with --chown on both
COPY layers so the app can still create/write bigregister.db (WP-22, a
relative-path SQLite connection string resolved against the container's /app
cwd) as that user.
Verified for real: rebuilt, confirmed `whoami` is `app` inside the container,
ran it and curled a live GET /api/v1/brief/preview (200), confirmed
bigregister.db was created and is actually owned by app:app. Full semgrep
re-run (this file didn't exist during the original triage) is now 0 findings.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Records what's implemented (items 1/3/4/5/6), what's deliberately skipped
this round (item 2, blocked on act_runner access), and that the WP can't be
marked fully done until a real Gitea run confirms the CI-timing/path-filter
behavior this environment can't observe. npm run ci confirmed green locally.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Multi-stage backend/Dockerfile (sdk build -> aspnet:10.0 runtime, ~312MB) +
docker-compose.prod.yml, additive only — not wired into CI or the existing
dev docker-compose.yml (which keeps the SDK image for dotnet run hot-reload).
New .dockerignore keeps the build context lean (node_modules alone is
~750MB) since the Dockerfile COPYs from the repo root to pick up
public/letter.css (WP-25's FE<->BE letter contract) as a sibling of backend/.
Verified for real: built the image, ran it, and curled a live
GET /api/v1/brief/preview against the running container — got back the
actual rendered letter HTML with letter.css inlined, confirming the
walk-up-from-BaseDirectory lookup resolves inside this image layout too.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
New `changes` job (dorny/paths-filter, pinned to SHA) computes frontend/
backend outputs; every downstream job gates its real steps on the relevant
output(s) instead of being skipped as a whole job. Conservative "skip steps,
not jobs" variant: every job still runs and reports a status (checkout always
executes) even when its side is untouched, so a required-status-check never
waits on a job that never started — the tradeoff the WP itself flagged as the
open risk of this item. e2e/semgrep/api-client-drift gate on either side
(they exercise both). `.github/workflows/**` counts as both sides, so a CI
change always gets a full run. Validated with `actionlint` (0 issues) and a
local YAML parse; the actual skip behavior can only be confirmed on a real
Gitea PR run.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
New `lint` job (lint + format:check + check:tokens) runs in parallel with
`frontend`, reporting in ~5 min instead of waiting on the full
test:coverage/ng build --localize/npm audit chain. Depends on the item-1
node_modules cache (otherwise this would duplicate a full npm ci for no
speed benefit, per the WP's own note).
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
actions/cache on node_modules keyed by package-lock.json hash, across all 4
npm-based jobs (frontend, storybook-a11y, e2e, api-client-drift). The if:
guard is what makes this a real speedup: npm ci deletes-then-reinstalls
unconditionally, so caching node_modules alone does nothing unless the
install step is skipped outright on a cache hit.
Unverifiable from this environment — CI timing needs a real Gitea run to
confirm; shipped as its own commit so a red run is easy to bisect/revert.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Local semgrep run found 25 findings (not the WP's remembered 27 — already-stale
by the time this ran): dependabot cooldown, npm min-release-age, every GitHub
Action pinned to a full commit SHA (dependabot's existing github-actions
ecosystem entry keeps these current), and 2 detect-non-literal-regexp findings
in e2e/create-ssp.mjs suppressed as false positives (non-attacker-controlled
input: a test's own captured version number, a local generator's CLI arg).
`semgrep scan` now runs with `--error`, a real blocking gate instead of
report-only.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Opt-in docker-compose (postgres+redis+OpenZaak, no celery/nginx) +
bootstrap-catalogus.sh seed a real OpenZaak instance; OpenZaakIntegrationTests
(Category=Integration, excluded from default dotnet test/CI) proves the ZGW
seam against it for the first time. That live run caught a real bug:
ZgwHttpClient never sent Content-Crs/Accept-Crs headers, so every write would
412 against a spec-compliant OpenZaak — fixed alongside the harness.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Replaces the hardcoded DocumentStore.DemoOwner and the static ZgwOptions
UserId/UserRepresentation with one per-request CallerIdentity, resolved by a
pluggable IIdentityProvider (StubIdentityProvider reads X-Role/X-Subject
today; a real OIDC/DigiD provider swaps in without touching any consumer).
- Domain/Authorization/{CallerIdentity,IIdentityProvider,StubIdentityProvider}.cs
+ a resolution middleware in Program.cs, right after correlation-id.
- Authz.ResolvePrincipal(ctx) keeps its signature (now reads ctx.Caller().Role),
so its ~15 call sites needed no changes.
- Every endpoint that passed DocumentStore.DemoOwner to a store now passes
ctx.Caller().Bsn.
- ZgwTokenProvider gains Mint(CallerIdentity) alongside the original Mint()
(kept for calls not tied to one citizen); ZgwHttpClient threads an optional
caller through to pick the right overload.
- IZaakSource gains ListMyCases(caller, now) — the citizen-scoped read
OpenZaakZaakSource backs with ZGW's rol__...__inpBsn filter. GET /applications
now routes through it instead of ApplicationStore directly, closing the last
"reads a static store" gap for a citizen-facing endpoint.
Backend 159/159 tests (+8, incl. an HTTP-level two-identity scoping proof),
npm run ci green, no api-client drift.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Endpoint/DTO/options landed already in c4dd846; this closes the loop with
NotificatieTests.cs (accept/reject/missing-header, asserting the AuthzAuditStore
row), missing appsettings.json keys (also backfills DrcBaseUrl/
InformatieobjecttypeUrls, stale since WP-51), and the webhook + abonnement
provisioning docs.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
.storybook/main.ts: webpack core's auto-DefinePlugin (from optimization.nodeEnv)
and Storybook's own preset both define process.env.NODE_ENV on the same
compile, triggering a "Conflicting values" warning even though both resolve
to "development" locally. Disable the redundant one via webpackFinal.
ci.yml: revert the storybook-a11y container's memory cap 6g -> 4g. The 6g
bump was based on an unconstrained local RSS measurement (~5.8GB) that
doesn't reflect real behavior under a cgroup cap. Verified directly: running
this job's exact steps (npm ci, playwright install, build-storybook,
test-storybook:ci) in `docker run --cpus=2 --memory=4g --memory-swap=4g
node:24-bookworm` completes clean, no OOM, 62/62 suites passing.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
- backend: dotnet format the WP-51 migration (2-space indent, no BOM)
to match .editorconfig — dotnet format --verify-no-changes was failing.
- storybook: stub FeatureFlagStore (WP-47) in shell/site-header stories
alongside AccessStore, fixing NG0201 no-provider errors; bump the
storybook-a11y container's memory cap 4g→6g (build-storybook +
compodoc measured ~5.8GB peak RSS, leaving too little headroom).
- backend: fix a startup-breaking bug in the new (WP-52) POST
/zgw/notificaties handler — it took ZgwOptions as a minimal-API
parameter, which isn't registered in DI, so ASP.NET's endpoint-table
build threw on every request once the route was registered (incl.
/swagger, which is why Playwright's webServer health check timed
out). Close over the existing `zgw` local instead.
- e2e: brief-v2.spec.ts's "Voorbeeld" button locator was ambiguous
once a second "Voorbeeld met testwaarden" button existed (Playwright
name matching is substring-based) — added `exact: true`. Also fixed
the sent-letter preview flow to match app-letter-composer's actual
behavior (single click → fetch, no in-page dialog, unlike
app-behandel-scherm's), and fixed a watermark assertion that checked
for the always-present `.preview-watermark` CSS class name instead
of the conditionally-rendered "VOORBEELD" marker text.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Every `npm run storybook` start ran a full Compodoc TypeScript-program
analysis over the whole app (not just the 62 stories), regenerating a
9MB documentation.json and spiking memory/CPU enough to crash the
machine. build-storybook (what CI/docs actually ship) keeps compodoc;
the interactive dev loop just serves the last-committed
documentation.json.
Extends the OpenZaak seam with IDocumentSource, sibling of IZaakSource
(WP-49/50): an upload always lands locally first (DocumentStore stays
the record of truth for preview/download/audit) and, when
Zgw:Enabled=true, is also registered as a DRC enkelvoudiginformatie-
object; once a zaak exists (IZaakSource.CreateZaak now also returns
its ZaakUrl), submit links each document to it via zaakinformatie-
object. FE upload/list DTOs are unchanged.
- ZgwOptions gains DrcBaseUrl + a category->informatieobjecttype URL
map (the document analogue of ZaaktypeUrls).
- LocalDocumentSource is the same DocumentStore.Add/Link calls the
endpoints used to make inline — zero behaviour change offline.
- OpenZaakDocumentSource POSTs the eio then the zaak link, persisting
the DRC url (DocumentStore.SetDrcUrl) so linking doesn't re-upload.
- Factored the GET/POST-with-bearer-JWT plumbing shared with
OpenZaakZaakSource into ZgwHttpClient; shared the stub handler
between the two source test classes as ZgwStubHandler.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Extends the IZaakSource seam (WP-49, read-only) with CreateZaak: submitting
an aanvraag now also registers a Zaak + Status + Rol in OpenZaak when
Zgw:Enabled=true, routed through the existing /applications/{id}/submit
endpoint with the FE response DTO unchanged (ADR-0001/ADR-0005 — the
endpoint never branches on the config flag itself, DI already picked the
implementation).
- ZgwOptions gains a Type→zaaktype-URL map + the two RSINs a Zaak needs.
- LocalZaakSource.CreateZaak is a pure passthrough of what the endpoint
already computes locally (zero behaviour change for the offline default).
- OpenZaakZaakSource.CreateZaak POSTs the zaak (identificatie = the same
local reference, so both stay in sync), resolves + POSTs the initial
status and the initiator rol (BSN) via Catalogi lookups, and maps the
result back into the submit response.
- Marked ponytail shortcuts: first-statustype/roltype-Catalogi-returns
(no per-type config) and no compensating transaction on partial failure
— both fine for a first slice against a demo backend.
Verified: full `npm run ci` green, zero api-client drift, 144/144 backend
tests (142 existing + 2 new stub-handler tests asserting the POST bodies
+ type→zaaktype mapping per the acceptance criteria).
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Mechanises the new-ssp skill's manual fork-and-strip recipe into a runnable
script: strips the four business contexts and every wiring point that
references them (routes, tsconfig aliases, dependency-cruiser boundaries,
nav links, capability union, the debug-state panel, showcase snippet
tooling), renames BigRegister.* throughout including the backend
solution/projects, re-runs gen:api, and reuses gen:context (WP-44) to seed
the new portal's first context. Backend business rules and real branding
can't be generated from nothing, so those print an explicit checklist
instead of pretending to be automated.
Verified end-to-end against an isolated clone (not this working tree):
lint, dep:check, check:tokens, all frontend tests, ng build, dotnet
build/test on the renamed backend, and gen:api against it all pass clean
with zero remaining references to the stripped contexts.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
The shell (and this switcher within it) is a persistent parent, so the
switcher's computed link only ever ran once at bootstrap against the
initial location.pathname. Navigating client-side afterward left it
frozen on that first route (typically /login), so switching language
from any other page sent you to the stale /en/login instead of the
current route — indistinguishable from being logged out, though the
session was untouched. Recompute on every completed Router navigation,
same toSignal(router.events...) idiom already used by the breadcrumb
in site-header.component.ts.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Closes phase 6 (Brief v2): a demo script mapping shipped scenarios to
URL+click paths (no Brief v2 PRD ever existed to translate one from —
written directly against the code instead), one e2e spec covering
compose→approve→send and admin republish→drafter-sees-it, and
Storybook state gaps (rejection diff, read-only viewer, org logo,
upload rejection) that prior WPs left uncovered. Flags passage-picker
as dead code, superseded by besluit-panel.
npm run e2e is not verified green in this sandbox — see WP-28's
Deviations section; the pre-existing, unmodified smoke.spec.ts fails
identically here, pointing at a sandbox rendering issue rather than a
regression.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Rewrite planner/developer/task-runner descriptions with explicit "use
proactively" triggers and the model-mismatch condition, so the
orchestrating session self-routes to the right agent instead of relying
on CLAUDE.md prose being remembered each time.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Pin the model to the step, not the whole session, via .claude/agents/*.md frontmatter
instead of relying on a human to run /model at the right moment: planner (Opus, no
Edit/Write) for design work, developer (Sonnet) for implementation, task-runner (Haiku,
no Edit/Write) for mechanical read-only checks. CLAUDE.md documents the routing rule;
the backlog session protocol now names these agents instead of a manual /model instruction.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
npm run gen:context scaffolds a bounded context: folders + starter page, the @<ctx>/*
tsconfig alias, a dependency-cruiser boundary entry, and a lazy authGuard route.
Refactors .dependency-cruiser.js's per-context contextRule calls into a single
CONTEXT_ALLOWED map that every rule derives from, so adding a context is really one
config entry (verified behavior-preserving: same dep:check counts, same graph output).
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
WP-27's Decisions said undo/redo is a hand-rolled past/future: Brief[] stack in
BriefStore. WP-31 later extracted that into the shared createHistory<T> helper
(the backlog's own phase-7 ordering note says as much: "32 reuses the
createHistory helper 31 extracts") — BriefStore now calls .record/.undo/.redo on
it, same cap-50/push-on-edit/clear-on-new-edit semantics. Noted the supersession
rather than leaving the WP describing code that no longer exists verbatim.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
WP-19: the Files section still described the originally-planned CI approach
(background servers with & + wait-on); the actual implementation (documented in
ci.yml's own comment and playwright.config.ts) moved server lifecycle into
Playwright's webServer to fix the 2-hour e2e hang. Recorded as a Deviation entry
and corrected the Files section to match reality. Spot-checked the rest of the WP
(retryClicked wiring, the currentScenario() claim, the still-open retry gaps in
registration-detail/aanvraag-detail/brief) against current code — all still hold.
Session protocol: replace the removed model mandate with per-step guidance —
Opus for planning, Sonnet for implementation, Haiku subagents for simple
read-only CLI checks — so cost matches the step's difficulty, not the whole WP.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Session protocol shouldn't dictate which model to use — WPs (including the new
WP-53/54) are written to be executable regardless of model.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
The two highest-value OpenZaak roadmap gaps, each written self-contained (a "current
state" handoff section) so a fresh session can execute from the file + repo alone:
- WP-53: replace the stubbed owner/BSN with a real per-request CallerIdentity
(pluggable stub, not DigiD), threading it into Authz, the ZGW JWT user claims, and
a citizen-scoped read (rol__…__inpBsn). Production-blocking for a real deployment.
- WP-54: a separate docker-compose OpenZaak + scripted bootstrap + opt-in
Category=Integration test — makes 50/51/52 developable against a live instance
instead of only fixtures; kept out of the default gate.
Indexed in the backlog README (rows + phase-9 ordering note) and cited from
openzaak-integration.md.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
- test-storybook:ci gets --maxWorkers=2 so the Jest runner stops spawning one
headless Chromium per core and OOM-ing the Gitea runner host (the root cause).
- storybook-a11y job gains a container resource ceiling (--cpus=2 --memory=4g) as
a belt-and-suspenders guardrail; noted it needs a docker-mode act_runner.
- openzaak-integration.md: add "Anti-corruption layer — two nested boundaries"
teaching section (BFF ACL vs upstreams + FE ACL vs BFF, and the principles).
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Stamdata: add beroepen, opleidingen (temporal), and specialismen tables to the
schema-driven catalog (zero UI code). opleidingen.beroep and specialismen.beroep
both reference beroepen.code — the first stamdata->stamdata references, enforced by
two new StamdataRef entries in the CI gate.
OpenZaak/ZGW (WP-49, slice 1 — read-only zaken): introduce IZaakSource as the cases
read seam. Default LocalZaakSource reads the local SQLite store (offline); an
OpenZaakZaakSource (Zgw/ client: HS256 per-call JWT, ZGW->existing-DTO mapper,
paginating HTTP source) is selected behind Zgw:Enabled (default false). The FE never
changes — same ApplicationSummaryDto, no api-client drift. Unit-tested with fixtures
+ a stub HttpMessageHandler; no live OpenZaak needed.
Docs: ADR-0005, reference/openzaak-integration.md, WP-49..52 roadmap, stamdata.md
update, README index rows.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
- Index roles-and-access, dependencies, and dependency-graph (were on disk
but missing from the docs README index).
- New reference/scaffolding.md: plop generators, gen:api (NSwag), gen:snippets,
and the skill recipes — the one consolidated "how generation works" page.
- Fix stale backlog range (WP-01…WP-28 -> WP-48) and de-link the dead WP-32 row
(undo folded into WP-31, no separate file).
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Three how-it-works/how-to-extend reference docs (docs/reference/), each with a
coupling section, indexed in docs/README.md. New document-feature skill so docs
ship in the same diff as the code.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
CI gate (authoritative): generalize the dangling-reference test in StamdataValidationTests
into a declared, extensible reference list ("every declared reference into a stamdata key
resolves against the currently-valid stamdata"), starting with Diploma.Opleiding →
professions.program. Removing/renaming a referenced program, or expiring it while current
data still references it, fails the PR build (ADR-0004). Editor (fast feedback): confirm
before delete (warns a referenced row fails CI) and, for temporal tables, a "Sluiten per
vandaag" action that closes validity (geldigTot) — steering to expire over hard delete.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Catalog declared in code (Domain/Features/FeatureFlags.cs, build-validated), on/off state
persisted in SQLite (FeatureFlagStore + migration). GET /flags (drives FE gating) + admin
PUT /admin/flags/{key} (new flags:manage capability + FlagsAdmin gate). Enforced end-to-end:
the `inschrijving-open` flag hides the Inschrijven nav item + dashboard action (FE) AND makes
POST /applications for a registratie 403 when off (backend). FE FeatureFlagStore mirrors
AccessStore (enabled() deny-by-default); admin toggle page at /beheer/functies in ADMIN_LINKS.
+4 backend tests, /me cap-list updated, client regenerated.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
The language switch is a full-page navigation to a separate bundle (nl at /, en at
/en/); sessionStorage's per-tab semantics dropped the login across it. localStorage
is unambiguously shared same-origin and survives the hard navigation. Keeps G1 (naam
only, never the BSN). Trade-off: the demo session now survives tab close — a real
portal keeps auth in an httpOnly cookie/token.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Serve the source locale (nl) at / instead of /nl/: angular.json sourceLocale is now
{ code: 'nl', subPath: '' } → nl output at browser/ root (base href /), en stays /en/.
Rework localeLinks (nl → bare path, en → /en/…) + spec, and serve-i18n.mjs (nl assets at
root, en under /en/, / serves the nl index). And build the docker demo (+ serve:i18n) with
`--configuration development --localize` so isDevMode() stays true and the dev `⚙ state`
panel + role/scenario switchers render in the localized compose demo (they were correctly
gated off in the previous production build). Verified: nl base href /, en /en/, ngDevMode present.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Runnable `npm run gen:value-object` / `gen:form-machine` (plop) that scaffold the two
pure-TS house patterns with a co-located spec: a branded value object + parseX (mirrors
postcode/bsn), and an Elm-style form/wizard machine (Draft/Valid/Errors + Editing/
Submitting/Submitted/Failed union + initial/pure reduce/assertNever). Prompts take
context + PascalCase name (positional-arg bypass); a post-action reminds to add the
English target for the generated $localize id. Templates in plop-templates/ (prettier-
ignored). Skills (value-object, form-machine) point at the generators. ui-component +
bff-endpoint stay skill-driven (Angular {{}} / backend + gen:api).
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Fix the hardcoded LOCALE_ID: 'nl' — provide it from $localize.locale (the build-time
locale, 'nl'/'en', undefined→'nl' in dev) and register both nl+en locale data, so the en
bundle formats dates/numbers correctly. Make `docker compose up` serve the LOCALIZED build:
the web service now runs `ng build --localize` then serve-i18n.mjs, which gained a PORT env
+ an /api reverse-proxy (API_PROXY_TARGET → the api container) so both /nl/ and /en/ are
served with the language switcher working end-to-end. Drop the now-unused proxy.conf.docker.json
(serve-i18n proxies /api itself); update ARCHITECTURE. `npm start` stays the nl-only HMR loop.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Add a language switcher matching the CIBG Taal-instellen pattern: a <nav> region
(sr-only heading + aria-label) with an endonym link per locale (lang/hreflang, the
active one aria-current + non-link), mounted right after the skip link in the shell.
Compile-time $localize means each locale is its own bundle under /<locale>/, so the
switch is a full navigation to the sibling bundle — active locale read from the baked
base href. Pure localeLinks() (+spec) builds path-preserving targets. Since `ng serve`
serves nl-only at /, add `npm run serve:i18n` (localized build + a tiny static server
with per-locale SPA fallback) so the switch is demoable. +story.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Surface the besluit-driven assistance that was previously silent: a pure
besluitGuidance() (kern passage count + needs-reason flag) rendered as a hint below
the besluit panel in behandel-scherm (warning to pick a reden, else info on how many
standaardteksten were inserted). Add a read-only notice on letter-composer for a pure
viewer (no edit/approve/reject/send right, e.g. admin) so the read-only letter isn't
mistaken for a broken editor — the "reverted to a view of the letter" confusion was a
role issue (non-drafter sees the composer). +besluitGuidance spec.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
The WP-33 panel's <select [value]> was applied before the @for options existed, so
the matching option was never selected (dropdown always showed the first item).
Move selection to [selected] on each <option>, and read role/scenario via getters
(currentRole()/currentScenario()) so the dropdowns reflect the active value whenever
the panel opens.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
The WP-41 GET /admin/audit trail now has an FE view: a beheer audit page (domain
AuditEntry + adapter/parse + store) rendering the data-minimised trail as a read-only
table, capability-gated on cases:manage. Added to ADMIN_LINKS (header nav + dashboard
Beheer section) and to the role.interceptor ROLE_AWARE list so the admin-gated call
carries X-Role. Closes WP-42's audit half.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Persist the security-relevant events (authz denials + BIG-nummer reveal/step-up) into a
data-minimised EF table (AuthzAuditEntry: At/Action/Resource/Decision/Role/CorrelationId —
never a name/BSN/value), extending the DocumentStore AuditEntry pattern (migration AuthzAudit).
AuditAuthz now persists via AuthzAuditStore.Record alongside its log line. GET /admin/audit
(admin-gated by the existing CasesAdmin) returns the trail newest-first. +3 backend tests incl.
a schema-carries-no-PII reflection test. Typed client regenerated (audit() + AuthzAuditDto);
no FE consumer yet (a future audit view must add the ROLE_AWARE prefix). Finishes WP-42's audit half.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Add a "PII — maskeren & parsen" section to /concepts demonstrating the WP-40 pieces
with FP + atomic design, framed for AVG art. 9 / data-minimisation: a live
<app-masked-value> atom (masked BSN that reveals on click; real reveal is step-up +
audited in behandel-scherm) and a live parseBsn elfproef parse. Both show the real
linked source via the WP-39 snippet mechanism (new showcase regions in bsn.ts + pii.ts,
registered in gen-snippets.mjs). Delivers WP-42's showcase demo; the persisted-audit
half stays pending WP-41.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Branded Bsn value object with the elfproef (11-test) checksum in shared/kernel/bsn.ts,
wired into the DigiD login boundary so login does real BSN validation (hint + e2e BSNs
updated to a valid 123456782). Consolidate the pure maskers into shared/kernel/pii.ts
(maskBsn/maskTail/REDACTED); debug-state keeps redactProfile (needs the registratie
BigProfile — boundary). New <app-masked-value> atom (+story) centralises the masked
`.includes('*')` detection + reveal affordance; behandel-scherm refactored onto it.
Session.bsn stays string (persistence boundary drops it for privacy). +specs for bsn/pii.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Anti-drift snippets: `?raw` isn't supported by Angular's esbuild build, so real code
is exposed via // #region showcase:<name> markers in source (registration/remote-data/
postcode/change-request.machine/intake.machine) → scripts/gen-snippets.mjs → committed
snippets.generated.ts, with a CI drift gate so the shown code is the shipped code. The
/concepts page renders the 5 real snippets (union/fold/parse/machine/steps) with a source
caption + a tiny dependency-free highlightTs (+spec); deliberately-wrong illustrations stay
authored. Teaching motion: active state node pops on transition + parse result animates in
(existing reduced-motion-safe toolkit). Generated file prettier-ignored (like api-client).
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Enable coverage via the @angular/build:unit-test builder's first-class options: a
`coverage` configuration on the test target (v8 provider, text-summary/html/lcov,
excludes for specs/stories/contracts/generated client) + `npm run test:coverage`.
Report-only (no thresholds) — visibility first. CI's test step now runs test:coverage
(strict superset: fails on a failing test, never on coverage). Baseline ~71% stmts.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Adopt dependency-cruiser as the single declarative source for bounded-context +
atomic-layer boundaries, replacing the per-context no-restricted-imports blocks that
had to be hand-copied (and had left herregistratie uncovered). `.dependency-cruiser.js`
encodes context direction (everyone→shared, herregistratie→registratie, showcase→*),
domain-purity, contracts-import-nothing, ui↛infrastructure, ApiClient confinement, and
no-circular. `npm run dep:check` enforces (wired into ci-local.sh + the frontend CI job);
`npm run dep:graph` emits a committed mermaid context×layer graph. ESLint slimmed to
no-explicit-any + template a11y. Docs + new-context skill updated to the single source.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
currentScenario()/currentRole() read the URL param before sessionStorage, so a
stale ?scenario=/?role= in the address bar overrode the switcher on reload
("stuck on slow"). The switcher now strips both dev params from the URL
(pure stripDevParams + history.replaceState) before reloading, so the stored
value wins.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
New phase from the 5-item analysis: linked showcase snippets + animations,
runnable scaffolding generators (pages/wizards/context/SSP), dependency graph +
declarative boundaries, PII hardening (branded Bsn VO + masked-value atom +
persisted no-PII audit) doubling as a privacy showcase, and Vitest coverage.
Goal: both a teaching showcase and a production starter template. Sequenced with
priorities + dependencies in the README.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Add docs/reference/roles-and-access.md — a short, scannable reference for the
three dev roles (drafter/approver/admin), that login is faked and role is a
separate dev-only stand-in, how to switch (dev switcher / ?role=), a per-role
capability table, and the one ABAC principle (server emits + enforces; UI
renders). Points to PRD-0002 + the code. Adds a pointer from CLAUDE.md.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Add a capability-gated Beheer section to the dashboard listing the admin pages
(Huisstijl, Stamdata, Aanvragen), each shown when the principal holds its
capability — so admin pages are discoverable, not URL-only. Extract the admin
link list to shared/layout/admin-links.ts as one source of truth, reused by the
site header (was ADMIN_NAV_ITEMS) and the new section. Capability-gated, never
role-derived (PRD-0002).
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
The runner's semgrep install failed trying to replace Debian's apt-managed PyJWT
("Cannot uninstall PyJWT ... RECORD file not found"). --ignore-installed installs
semgrep's deps fresh without uninstalling the apt copies, staying within the
runner's constraints (no setup-python, no container job).
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Admin-only overview of all cases across owners + an admin delete, gated by a new
`cases:manage` capability (Authz role→cap + CanManageCases + CasesAdmin gate;
FE capability + guard + nav + role.interceptor prefix — the org-template/stamdata
recipe). Backend adds ApplicationStore.ListAll()/DeleteAny() and GET /admin/cases +
DELETE /admin/cases/{id}; admin delete removes ANY case incl. submitted. Page lives
in registratie/ui (owns the Aanvraag aggregate; reuses aanvraag-view + parse),
routed /beheer/zaken; delete guarded by a native confirm, optimistic with rollback.
Typed client regenerated (documents the new endpoints + owner field).
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Make "at most one unsubmitted Concept per type" a server invariant instead of a
client-only convenience. ApplicationStore.Create → CreateConcept guards atomically
under the write gate and POST /applications returns 409 when a duplicate would be
created. The FE draft-sync recovers from the 409 by adopting the existing Concept
(ensureId → findConcept) rather than erroring — one-per-type means the second
attempt lands on the existing draft. Typed client regenerated (documents the 409).
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Restyle the contact-change form with house CIBG surfaces instead of hand-rolled
markup: the read-only BRP address now uses app-data-block/app-data-row (grey
datablock), and the phone field sits in a plain <fieldset> inside the existing
.form-horizontal form (inherits the CIBG grey box). No new CSS surfaces; row keys
reuse the address.* i18n ids. Logic/machine/specs unchanged.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Reshape the adreswijziging form into a contact-change form: the BRP address is
authoritative and shown read-only (you change it at the gemeente), and the phone
number becomes the editable/submittable field. New Telefoonnummer value object
(parse-don't-validate); backend RejectPhoneChange re-validates as authority.
POST /change-requests now carries { telefoon } (typed client regenerated).
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Surface the ?scenario= and ?role= dev stand-ins as dropdowns in the existing
debug-state devtool, so a demo can flip them with a click instead of editing the
URL. scenario.ts/role.ts gain set* setters + exported valid-value lists (reused
by the panel, no duplicated source of truth); scenario becomes tab-sticky like
role so it survives navigation. Applied via location.reload() since both are
read per-request in interceptors. Extends the debug-state eslint exemption to the
ui→infrastructure rule (same devtool precedent).
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Wire the WP-31 createHistory helper into StamdataStore: per-table undo/redo
over the edited rows, recording only real edits and restoring via the existing
Seed msg. Ctrl/Cmd+Z / +Shift+Z, deferring to native text-undo inside grid
cell inputs.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Audit "apply high-value": extract four shared helpers into shared/application/ and
rewire the editor stores (behaviour unchanged, existing specs are the gate):
- action-state.ts: ActionState/SaveState (were duplicated in both brief stores).
- history.ts: createHistory<T> (extracted from BriefStore's WP-27 undo/redo; WP-32 reuses).
- debounced-save.ts: createDebouncedSave (the 600ms timer/PendingSave dance, was 2×+).
- machine-remote-data.ts: machineRemoteData (the loading/failed/loaded→RemoteData switch, 3×).
Each helper has a co-located spec. Deferred DDD findings (contracts/ inconsistency, a
parse* traverse combinator, the 6× Seed boilerplate) are reported in the WP file, not built.
npm run ci green; 323 tests (+13 helper specs); brief/org-template/stamdata specs unchanged.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
WP-25 (1bb9383) and WP-27 (5e36d68) were done but the table still said todo — fixed.
WP-28/WP-30 remain todo. Added Phase 7 (refinements & hardening): WP-31 shared store
helpers, WP-32 stamdata undo, WP-33 dev switchers, WP-34 adres phone + BRP read-only,
WP-35 one Concept per type, WP-36 admin cases page + delete. Only 31→32 is ordered.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
The container: approach failed — this act_runner times out pulling its base runner
image (docker.gitea.com/runner-images:ubuntu-latest, IPv6) for container jobs. And
the earlier setup-python step failed downloading Python. Both avoided: run on the
plain ubuntu-latest runner and install semgrep with the preinstalled python3/pip
(`python3 -m pip install --break-system-packages semgrep`; --break-system-packages
survives PEP-668, pip puts semgrep on PATH). Verified in a clean python:3.12
container that pip install lands `semgrep` on PATH and the scan runs. Still
report-only; WP-30 tracks the flip to --error.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
The setup-python + `pip install semgrep` step errored on the Gitea runner. Switch to
`container: docker.io/semgrep/semgrep` (semgrep preinstalled) — the documented way to
run Semgrep in CI, and the exact execution verified locally (306 rules / 450 files,
27 findings, exit 0). Fully-qualified image name so short-name resolution works under
Docker or podman. Still report-only (no --error); WP-30 tracks flipping to blocking.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
CodeQL is GitHub-only — its analyze step uploads SARIF to GitHub's code-scanning
API and assumes a GitHub Security tab; this CI runs on Gitea only, so the job could
never go green (it had been red since it was added). Replace it with Semgrep OSS, a
plain CLI SAST with no account/platform API, which runs fine on Gitea.
- Remove the codeql job (+ its security-events permission) and the schedule trigger
(it existed only for codeql; semgrep runs on push + PR).
- Add a semgrep job: setup-python + `pip install semgrep` +
`semgrep scan --config p/default --config p/csharp --metrics=off`. pip-on-runner
(not container:) mirrors the other jobs' model; anonymous registry, telemetry off.
- Report-only for now (no --error → job stays green): a local dry-run found 27
findings, mostly CI/config policy (unpinned actions, .npmrc), not app-code vulns.
WP-30 tracks triaging them + flipping to --error (a blocking gate).
Verified locally: `semgrep scan` runs clean (exit 0 without --error, 306 rules /
450 files). CI behaviour confirmable only on the Gitea runner — watch the run.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
The header now injects AccessStore (→ MeAdapter → ApiClient) for its admin links.
f719676 stubbed that in the site-header story but missed shell.stories.ts, which
renders the header via ShellComponent — with only provideRouter it threw NG0201
(no ApiClient provider), failing the Shell smoke-test and turning storybook-a11y
red on CI (deterministic, not the local worker-contention timeouts). Stub
AccessStore there too (can() → false, no admin links).
Verified: full test-storybook suite single-worker now 60 suites / 170 tests green.
Lesson: when a shared component gains a service dep, every story that renders it —
directly OR via a parent template — needs the provider.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
The admin pages (/beheer/stamdata, /brief/huisstijl) were unreachable in the browser,
for three compounding reasons — all fixed here:
- **Guard raced /me.** capabilityGuard read can() synchronously while /me was still
loading, so it denied even an entitled admin (deny-by-default) and bounced to /login.
It's now async: awaits AccessStore.whenReady() (new — resolves once /me settles), then
allows if entitled; an authenticated-but-unentitled user goes to /dashboard, anonymous
to /login. + auth.guard.spec (the missing test that let this ship).
- **Dev role wasn't sticky.** currentRole() read ?role= from the URL on every request,
but login/nav drop the param, silently reverting admin→drafter mid-session and 403-ing
the admin endpoints. It now persists the role per-tab (sessionStorage), so every
role-aware request keeps it. Dev-only (the interceptor is wired only under isDevMode).
- **No way in.** Added capability-gated Huisstijl + Stamdata links to the header (shown
only when /me grants the cap); injecting AccessStore there also warms /me early. New
en translations for the two labels; site-header story stubs AccessStore (+ AsAdmin
variant) so it needs no HTTP.
Verified live: with ?role=admin the header shows both links, clicking Stamdata loads the
grid (GET /api/v1/stamdata → 200, was 403→redirect); a non-admin sees no link. Full
`npm run ci` green (310 tests); site-header stories pass axe.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
The stamdata admin editor's adapter calls /api/v1/stamdata via the generated
ApiClient → roleInterceptor, but ROLE_AWARE omitted /api/v1/stamdata, so no X-Role
was sent and the backend StamdataAdmin gate resolved Drafter → 403 on every read
(confirmed: 403 without X-Role, 200 with X-Role: admin). Added /api/v1/stamdata to
the allow-list (same class of gap WP-23 fixed for /me) + a roleInterceptor spec so
the next admin endpoint isn't forgotten.
Note: a separate issue still blocks the page in the browser — capabilityGuard
redirects both admin routes to /login because it checks can() before /me resolves;
tracked separately.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Capture the deferred CI options from the speed pass (708d4c2) so they aren't lost:
node_modules cache to skip npm ci, smaller CI runner image (act_runner label→image,
needs admin; or repo-only container: for node-only jobs), path-filtered jobs, a fast
split lint job, and an optional lean prod backend image. Notes the constraints (CI not
observable from the agent env; compose images aren't used by CI). Added the README row.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
CI (Gitea Actions, .github/workflows/ci.yml):
- CodeQL off the PR critical path: runs on push-to-main + a weekly cron only
(`if: github.event_name != 'pull_request'`, `schedule: Mondays 03:00 UTC`).
The 2-language 20-min matrix was the slowest thing on every PR; code is still
scanned on main + weekly.
- Cache Playwright browsers (~/.cache/ms-playwright) in the storybook-a11y + e2e
jobs — skips the chromium download on a hit; `install --with-deps` then only does
the fast apt deps check.
- Cache NuGet (~/.nuget/packages, keyed on **/*.csproj — no packages.lock.json) in
the backend / e2e / api-client-drift / codeql-csharp jobs.
- `npm ci --prefer-offline --no-audit --no-fund` in the 4 npm jobs.
Demo (docker-compose.yml, local only — NOT used by CI): web image node:24 →
node:24-slim (~1.1GB → 232MB verified). The container only runs `npm ci && ng serve`
and the native deps ship prebuilt glibc binaries, so slim needs no toolchain —
verified: npm ci clean, ng serve boots, app returns 200.
Note: CI timing/behaviour can only be confirmed on the Gitea runner (not observable
locally). Every change here is independently revertable.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
The e2e job backgrounded `dotnet run &` and `ng serve &` in separate Actions
steps, then `npx wait-on` (no timeout) in a later step. A process started with
`&` in one step is killed when that step's shell exits, so wait-on waited forever
on servers that were already gone — the job hung until the runner's hard limit
(~2h; Gitea's act_runner doesn't reliably enforce timeout-minutes).
Move both servers into Playwright's `webServer` (an array: backend + `npm start`),
so Playwright starts them, waits for readiness, runs the suite, and tears them
down in the one `npm run e2e` process — CI and local alike. The CI e2e job is now
just npm ci / playwright install / npm run e2e. `reuseExistingServer` is on locally
(reuses a running app, incl. the docker stack) and off in CI (fresh start).
Verified locally via `CI=1 npm run e2e` against free ports + a clean db: both
smoke tests pass in ~15s (no hang). The earlier local failure was leftover SQLite
state (WP-22) from a dirty run resuming a Concept — CI checks out fresh, so its db
is always empty.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Three fixes so Gitea CI is green and stays that way:
- **ng build --localize**: `messages.en.xlf` (hand-maintained, matched by id) had drifted
~93 units behind the source since WP-24 — `beheer.*`, `brief.canvas.*`, `brief.diff.*`,
`orgTemplate.*`, `brief.besluit.*`, etc. With `i18nMissingTranslation: "error"` that fails
the frontend job. Added English `<target>`s for every missing unit (interpolation `<x/>`
tags preserved); both nl + en bundles build clean.
- **docker npm noise**: the web container's `npm ci` now runs `--no-fund --loglevel=error`,
so `docker compose up` boots without npm 11's deprecation / allow-scripts / funding
warnings (verified in the running container: 0 warning lines). The underlying Angular
webpack-builder deprecation is upstream; migrating to @angular/build is out of scope.
- **local CI check**: `scripts/ci-local.sh` + `npm run ci` mirror the CI jobs (lint,
format:check, check:tokens, test, build --localize, audit, backend format+test,
api-client drift; `--full` adds storybook-a11y) so a red build is caught before pushing.
Opt-in pre-push hook at scripts/githooks/pre-push. Documented in CLAUDE.md.
Verified: `npm run ci` passes end-to-end; storybook-a11y 169/169 (the local parallel-worker
timeouts were the docker stack starving chromium, not a regression).
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
`npm run format:check` (a CI gate) had drifted red across 44 files — pre-existing
files plus recently-added ones committed without formatting. Ran `prettier --write .`;
no logic changes. Also regenerates documentation.json (compodoc reflects the reformatted
component sources).
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Rewrite all five mermaid blocks to portable syntax so they render (verified via
a mermaid-native render): split two `classDef`/`class` statements that shared a
line (the "why not just signals" graph); drop the `reduce() — PURE` participant
alias and the `;` inside a Note (§2c sequence); remove `<br/>` and parenthetical
transition labels from the two stateDiagram-v2 blocks (RemoteData, intake),
moving the aside into a note; normalize `<br/>`→`<br>` and drop `<b>` in the
atomic-hierarchy flowchart.
Extend §2g to document the now-closed gap: a CanDeactivate guard flushes the
pending debounce before in-app navigation, and beforeunload warns + best-effort
flushes on hard close (noting the HttpClient/keepalive limitation).
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Close the last-mile autosave gap: a debounced edit made in the final <600ms
before leaving a page was lost — the wizard draft-sync timer is cleared on
destroy without flushing, and root stores keep an armed timer the teardown
ignores.
New `shared/application/pending-saves.ts`: a root `PendingSaves` registry every
autosave owner joins (BriefStore, OrgTemplateStore, each createDraftSync). Two
seams flush through it — `flushPendingGuard` (CanDeactivate, on the five
autosave routes) awaits the pending write before an in-app route change; a
`beforeunload` handler (provideUnloadFlush) fires it best-effort and raises the
browser's native unsaved-changes prompt. ponytail: the HTTP seam is Angular
HttpClient (no keepalive/sendBeacon), so a hard-close flush can't be guaranteed
— hence the prompt; upgrade path noted in a comment. Each owner now nulls its
timer handle on fire so `hasPendingSave()` is accurate, and exposes
`flushPending()`.
Verified live against the running stack: navigating away 91ms after a keystroke
(well inside the debounce) fires one PUT /brief before the route changes; a
dirty reload raises the prompt, a clean reload does not. FE lint / check:tokens
/ 299 tests (+11) / build / build-storybook green.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
ehoandClaude Sonnet 5