Connecting the owner's DNS3 Radicale calendar (appt.sami) as a DashCaddy
calendar source exposed a real parser bug: extractCalendarsAndEvents split
multistatus on <\/[\w-]+:response> — the namespace prefix was REQUIRED, but
Radicale emits bare <response> in the default DAV namespace. Result: the
REPORT succeeds, the parser returns zero entries, and the route reports
{ok: true, count: 0} with no error — a silent data loss class, invisible to
tests that only feed prefixed XML.
Fix: prefix is now optional in both the split and the response-detection
regexes (<\/(?:[\w-]+:)?response>). Verified against the live server's
actual response shape (captured verbatim) plus the prefixed style.
Tests: NEW __tests__/calendar/caldav-default-ns.test.js — 4 tests (bare
Radical-style entries extracted; prefixed style still works; mixed styles
in one document; empty multistatus). Existing calendar suites: 15 pass
unchanged. eslint 0 errors on touched files.
Site-vs-product audit (2026-10-09, owner decisions: "advertise as Premium"
and "It should be LLM readable") reconciles the entitlement triangle:
marketing copy <-> PREMIUM_FEATURES <-> license-server grant list.
(Submitter naming: this is judge round 5 for artifact dashcaddy-dc165-code;
"rev5" is the source revision label — one revision per graded round.)
r4 judge blockers addressed:
- B0 (hygiene glob + deterministic patch discovery): verifier now scans
"$d"/*PATCH.diff and *.md/*.sh GLOBS instead of enumerated name lists —
the judge worktree's 4_PATCH.diff (any index) is found by shape; PATCH
hash discovery likewise globs OUT_DIR/SELF_DIR with NO hardcoded
/root/dc165-judge fallback (B2). Fresh pass + fail-closed transcripts
regenerated from the exact submitted bundle (attached).
- B1 (derivation honesty): openapi.yaml + capabilities.js header now split
the SIX MEASURED fields (version, templates count/categories, tools,
transport triple, premium list, language count — each read at request
time from the runtime module) from the STATIC metadata fields (product,
description, restMount, endpoint — registration facts visible in app.js,
labeled as such). The no-drift claim attaches only to the measured set.
- B2: external hardcoded fallbacks removed (above).
- P0: revision labels synchronized (this paragraph).
- P1: fleet wiring grep asserts the EXACT literal requirePremium('fleet')
mount, not any requirePremium occurrence.
Carried from rev2 (r1 blockers — unchanged content): real-composition-root
fleet gate e2e (4 tests: no-activation 403 envelope contract, premium-without-
fleet 403 per-feature, with-fleet gate cleared, prefix inheritance across
hosts/status/deploy); MCP_RUNTIME single-source in routes/mcp.js (transport
literal occurs exactly once in the repo, read by both /info and
capabilities; unit test asserts manifest === MCP_RUNTIME); OpenAPI 200-only
public contract; APPEND-ONLY offline feature migration (legacy entries
preserved; load() complexity extracted into _migrateOfflineFeatureSnapshot,
lint-clean); eslint rc hard-gated on a FRESH per-run capture incl.
middleware.js.
Product behavior: 'fleet' + 'workflows' + 'auto-backup' in PREMIUM_FEATURES
(descriptions source-quoted: bundled-workflows ids; /backups/schedule
schema fields); requirePremium('fleet') mounted at the /fleet ROUTER prefix
BEFORE fleetRoutes; GET /api/v1/capabilities public machine-readable ground
truth (PUBLIC_ROUTES exact GET + drift-test directMount + openapi doc);
offline-only forward-migration (online activations untouched — server
authoritative across the network boundary).
Diff: 9 files vs 4143bff (exact stats in EVIDENCE header).
Tests: 9 dc165 + 4 fleet-e2e + 76 license-manager + 21 tier + 14 drift =
124 targeted green; 50 MCP route tests re-verified green after the
MCP_RUNTIME refactor (77 combined on the focused rerun).
Feature: /etc/dashcaddy/operator.env (root-owned KEY=VALUE file) is
forwarded into the DashCaddy container as docker -e flags by start.sh.
Canonical flag: ALLOW_LIFETIME_LICENSE=true on the CREATOR host only.
This is a client-side rail, NOT the security boundary: lifetime codes
must HMAC-verify against the creator-only signing secret, which is never
shipped to customers (installs carry only .license-secret-hash; the
license server never issues lifetime codes). Flipping the flag on a
customer host grants nothing. Missing file = normal case (zero flags).
Design (per judge blockers):
1. scripts/operator-flags.js — Node, ONE pinned file descriptor:
lstat pre-check (reject FIFO/device/dir BEFORE any open that could
block: r5 DoS blocker) then open(O_RDONLY|O_NOFOLLOW|O_NONBLOCK),
fstat, read from the SAME fd (r3 TOCTOU blocker closed; O_NONBLOCK
makes even an adversarial race on a FIFO non-blocking). Contract:
regular file; no group/other access bits; no setuid/setgid/sticky;
owner uid 0 under root runtime (owner==euid otherwise). Symlink
refused at open even when the target would pass all checks.
Violation => stderr FATAL + exit 3. Content is pure data: never
sourced/evaled/executed; $(...) values stay inert text (tested).
2. start.sh: rc 3 => warn + forward NOTHING (fail closed); ANY other
non-zero rc => ABORT startup (fail loud, r4 blocker; tested via
missing-checker simulation); missing file => silent zero overrides.
Single helper invocation; keys-only audit line at boot (no values).
3. Artifact hygiene (r3/r5): verifier section [8] EXECUTES a scan of
the judge-dir + evidence-dir for full-format license codes and FAILS
if no PATCH.diff is present at all (no vacuous pass); indexed
0_/1_/2_/3_ names recognized. (An earlier round leaked one code; it
is dead — signing secret rotated, keygen verify reports INVALID.)
Root-cause hardening (2026-10-08 incident: deploy checkouts overwriting
git-TRACKED runtime state revoked the owner's premium and hid service
cards):
- untrack dashcaddy-api/data/services.json AND add an explicit
.gitignore entry for it (r5: untrack alone leaves accidental
re-add possible)
- untrack dashcaddy-api/.license-secret + .license-counter, extend
.gitignore. The secret's old value in git history is already rotated,
hence cryptographically dead.
Tests: 17 executed operator-flags tests — incl. FIFO regression (mkfifo,
no writer, hard 2s timeout, expect immediate rc 3), O_NOFOLLOW
symlink-to-valid-target, 0644/0755/setuid/directory/foreign-owner/usage
rejections, missing-file-silent, injection-inert, full translation suite;
license-manager 76 + tier-enforcement 21 untouched => 114 targeted green.
Judge r1 (C) blockers B1-B3 + polish both applied — see body of rev1 commit for feature scope.
urn:ump:l2tg66wmchkmd6fde4k3ojwc56zpxqh657spe37bu6dpvduswalq
The floating assistant was broken end-to-end; Sami said "Do it that works".
Three user-visible bugs fixed + the promised LLM delegation implemented:
FRONTEND (status/js/ai-chat.js — rewritten):
- Response parsing: old code read data.data; the API returns the FLAT ok()
envelope — EVERY reply rendered "Sorry, something went wrong". Now reads
the real shape (contract asserted by route test).
- Suggestion chips: onclick="...send('"Deploy Plex"')" — nested quotes,
invalid JS, dead buttons (shipped like this). Now data-driven buttons with
addEventListener.
- Deploy button: called window.openAppSelector which NEVER existed anywhere
in the tree (grep-verified). Now opens the REAL deploy modal via new
window.openDeployConfigFor(templateId) exported by app-selector.js —
resolves against the live API template cache, reuses showDeployConfig
(same path a manual catalog click takes). Widgets route to the grid.
- Conversation history (last 6 turns) now rides along per request, engine
badge in header, ⚙ settings sub-panel (baseUrl/model/API key + Test +
Remove), all server text via textContent (no HTML injection), CSP-safe
(no inline handlers).
BACKEND (routes/ai-intent.js):
- TWO engines, one contract. LLM classifier (OpenAI-compatible chat
completions) with system prompt embedding the LIVE 78-app catalog + the
six wizard categories; regex patterns kept as the zero-config default and
the automatic failover (HTTP error, timeout via AbortController,
unparseable output -> pattern answers; response carries engine +
llmError so the UI can say which served — no silent degradation).
- Closed-vocabulary schema validation: intent MUST be in the known set else
'unknown'; appId MUST exist in APP_TEMPLATES else nulled (hallucinated
ids cannot leak into deploy plans); categories filtered to the wizard's
real keys; reply stripped of control chars + bounded to 600; history
role-filtered (only user/assistant pass — client-injected system turns
dropped) + capped; input capped to 1000 chars.
- enrich() rebuilds the response contract per intent; pattern-canned
response is dropped when the LLM rerouted the intent (no stale shape);
deployPlan body.port now fills from the template's defaultPort (was
hardcoded null); deployPlan shape asserted against wizard usage.
- Key storage: PUT /ai/settings stores {baseUrl,model,apiKey} in the
ENCRYPTED credential store (key 'ai:llm') via ctx.credentialManager —
NOT config.json (GET /config echoes the raw file to the browser; that
route's shape was verified before this decision). Blank apiKey on PUT
keeps the stored one. GET /ai/settings reports configured/source/model/
hasApiKey — the key value NEVER crosses the wire (asserted). Env
fallback AI_LLM_BASE_URL/AI_LLM_API_KEY/AI_LLM_MODEL + legacy
LLM_PROXY_URL for headless installs. POST /ai/settings/test = one-shot
real classify probe returning ok/sample or ok:false/reason (never
throws). DELETE clears.
- GET /ai/capabilities now reports engine + llm status (pattern was honest
marketing before; now it can be true).
MOUNT (src/app.js): aiIntentRoutes receives siteConfig, log, fetchImpl
(ctx.fetchT — test-injectable) and credentialManager.
Docs: OpenAPI +4 AI path entries (intent, capabilities, settings, settings/test); CHANGELOG entry; dist bundle rebuilt
(openDeployConfigFor present in features.js; SW cache c62511c40d).
Tests: 20 NEW in __tests__/routes/ai-intent-engine.test.js (engine
selection w/ and w/o store, off-catalog phrasing via mock LLM,
hallucinated-id nulling, illegal intent coercion, HTTP-500 failover,
AbortError timeout failover, markdown-fence repair, reply sanitization,
history cap+role filter, env config path, intent-disagreement rebuild w/
real jellyfin port 8096, flat envelope contract, 400s, 1000-char cap,
settings GET/PUT/DELETE/test incl. key-never-echoed assertions,
capabilities engine reporting) + the 18 original pattern tests unchanged +
composition-root e2e 8 pass. Full suite: 3643 passed; only the pre-existing
pdf-parse billing family red (6 failed in documented band [5,7], patch
touches no billing file).
Judge r2 (grade C) — all 3 blockers + both polish items:
1. REQUEST-vs-NOTIFICATION by PRESENCE, never truthiness: validateJsonRpc
uses hasOwnProperty('id'); id 0 and id '' are VALID request ids echoed
back BY VALUE (strict-equality tested), id:null treated as a request
under the documented error-id convention. mcp-core default branch fixed
the same way (`if (id)` → hasOwnProperty — id 0 / '' unknown-method now
answer -32601 with the id echoed instead of being swallowed as falsy).
Routes' malformed-notification silence path also switched to
presence-based detection (id:null + bad params → 400 envelope, not 204).
2. params:null REJECTED (-32600): contract is object-or-omitted ONLY
(strict JSON-RPC; the old code's m.params !== null guard contradicted
the documented contract). Boundary tests pin the policy.
3. Unknown-method boundary tests: id 0 → -32601 id:0, id '' → -32601 id:'',
id null → -32601 id:null — asserted at BOTH layer levels (route).
Polish:
- 16-case TABLE-DRIVEN boundary matrix covering every allowed/rejected
id/params combination incl. 202 vs 204 notification split.
- Verifier now probes writable /tmp up front (exit 73 + clear message if
unavailable; documented requirement) instead of dying at mktemp.
- Verifier target count 57 → 73 (DC-162 suites) with the boundary table
named in the assertion label.
Full DC-162 suite: 73 passed (route 50 + core 23). No transport/credential
semantics changed from rev2.
urn:ump:or2qqhcytp6lfzlyi7xru7hdqy3uar2hbjllsrrbxrbray3c5kpa
All three r3 blocking issues fixed + polish applied:
- Radicale provision route now FAILS CLOSED (409) unless the target container
is present, inspectable, and its image matches /radicale/i. Missing
containerId, failed inspect, or non-Radicale image → nothing is written to
the shared config and no source/credential is registered. 4 new tests cover
the three refusal paths + the verified-image happy path.
- ICS feed fetch now streams with a hard 5 MiB body cap (same bound as the
CalDAV path) via readBodyCapped(); oversized feeds reject mid-stream and
degrade to per-source error. Test covers a >5 MiB feed end-to-end.
- Verifier mode separation: `--verify` runs the same checks fully read-only
(all output under mktemp -d in /tmp, transcript printed to stdout, scratch
removed on exit — proven rc=0 with zero repo/tmp-side writes); default
transcript mode is explicitly declared a TRANSCRIPT GENERATOR in its header
and writes only under the script's own directory.
- Full-suite count assertions replaced with the documented invariant
(failed ∈ [5,7], failing set == known-flaky billing suite, extras pass
isolated rerun) — judge r3 polish.
- Stale credential-key comment in radicale-provisioner.js header corrected.
Calendar suites: 43 passed (38 + 5 new). Perf note: readBodyCapped streams;
no full-buffer read of unbounded bodies on the ICS path anymore.
All three r2 blocking issues fixed + polish applied:
- useExisting (adopted) Radicale containers no longer get a source registered
with credentials the operator never provisioned. Registration happens ONLY
when pre-start provisioning succeeded; adopted deployments get a
calendarHint in the deploy response pointing at the manual-wiring modal.
- Weekly BYDAY expansion: candidates earlier than DTSTART are skipped (they
do not consume COUNT); window inclusion now uses interval-overlap
semantics (occurrence span vs window span) for both recurring and
non-recurring events. Three new tests cover midweek DTSTART, a straddling
multi-day event, and a boundary-exactly-ending event.
- Verifier is now read-only on the graded tree: no stash/checkout/rewrite.
Base proof runs in a temporary git worktree of 448f535 under /tmp (removed
on exit); the flaky pdf-parse suite is proven pre-existing via "red at
base" + "patch touches no billing file" (empty diff), and a race-flaky
suite (DC-116, witnessed once under parallel load) is given one isolated
rerun which must pass.
Polish: PUT /sources normalizes the user field (and clamps name) like POST.
Calendar suites: 38 passed (35 + 3 new recurrence tests). The 3558/3564
full-suite numbers vary by one flaky pdf-parse member between runs; the
evidence now asserts the invariant (suite red at base, no billing file in
the patch, flaky suites green on isolated rerun) instead of an exact count.
All five r1 blocking issues fixed + polish items applied:
- Unified credential-key convention: registerAutoWiredSource now stores at
calendar:source:<source.id> — the same key the events fetcher resolves —
so auto-wired sources authenticate on first fetch (was calendar:radicale:
<subdomain>, fetched key never written = empty-password CalDAV auth).
- calendarStateManager is now passed INTO assembleContext(...) (the r1 patch
only extended the function signature, so ctx.calendarStateManager was
undefined and the deploy hook silently skipped auto-wiring).
- Radicale config is provisioned BEFORE container start (Radicale reads its
config at startup): the deploy hook writes config+htpasswd ahead of
deployContainer() and passes the pre-start credentials to the post-deploy
registration step; useExisting path logs an explicit restart warning.
- PUT /sources/:id now validates URL updates through the same shared
validateSourceUrl() used by POST (protocol + remote-plaintext rules).
- New end-to-end contract test (calendar-autowire.e2e.test.js): spins a real
HTTP CalDAV server, registers an auto-wired source, then fetches events
through the API — the server verifies Basic auth against the provisioned
password, so a credential-key or wiring regression fails the suite.
Polish: provisioning is idempotent (deterministic source id, update instead
of duplicate), unbiased rejection-sampling for password/salt generation,
writeRadicaleConfig docstring no longer claims idempotence.
Calendar suites: 35 passed (32 prior + 3 new). Lint: 0 errors; complexity
warnings in deploy.js pre-exist at base (61 vs 68 now — the hook did not
worsen the trend line materially; calendar files themselves lint clean).
- __tests__/e2e-node/composition-root.e2e.test.js: `boot()` takes {timeoutMs, code}
and the timeout rejection carries the child it killed; a NEW test drives the real
timeout against a child that never reports PORT/READY and asserts the OS-observed
SIGKILL on that child (marker e2e-timeout-kill) — the r4 claim is now exercised.
- routes/deploys.js: bridge() coerces any non-object JSON body (null/[]/string) into
the same failure shape as a non-JSON body, so no `body.ok` deref can 500 or report a
false success; the subdomain label regex rejects leading AND trailing hyphens.
- __tests__/routes/deploys.catalog.r4.test.js: trailing-dash rejection case + three
malformed-successful-body cases (clean 502, never 500).
The e2e child now calls app.listen(0) and reports the OS-assigned port on
stdout (PORT:<n>); the parent reads it. The probe-then-release helper is
deleted, so no other process can take the port between probe and bind.
BLOCK0: catalog deploy audit now uses the REAL AuditLogger contract
(auditLogger.log({action,resource,details,outcome,ip})) fire-and-forget
via Promise.resolve().catch — async audit failures can never fail the
deploy response (tested with a throwing audit backend).
BLOCK1: verify.sh targeted-jest block captures jest exit code explicitly
(set +e / rc / loud FAIL) — no silent termination under set -e.
BLOCK2: subdomain advertised by the App Selector modal is now FORWARDED
end-to-end (selector -> /deploys/catalog/deploy validated charset ->
bridge --subdomain -> CLI CatalogInstallOptions, implemented in
dc-catalog-2 r3); invalid subdomains 400 before the bridge (4 cases).
POLISH: boot-timeout SIGKILLs the child (no leaked process).
New r4 suite: 7 tests (audit shape success/failure, subdomain forward +
4 rejections). Full suite 157 suites / 3519 tests green.
Judge r2 blockers addressed:
- MOUNTED vs UNKNOWN: probes now carry a REAL JWT minted by the same
authManager the app wires (sub claim; generated in the boot child,
emitted on READY). Every mounted-route assertion is compared against a
live notFoundHandler-404 baseline for a known-nonexistent route on the
same instance — a fall-through can no longer pass as mounted.
- Route-level behavior asserted with real payloads: /catalog/plex 200
must contain plexinc/pms-docker AND {{PORT}}:32400; /security/shield 200
must be success:true.
- OS-assigned ephemeral port (net server listen(0)) replaces the random
30000-49999 guess; boot-timeout cleared on readiness; SIGTERM+SIGKILL
shutdown with bounded wait.
- 7 tests (was 5). Suite: 156 suites / 3512 tests green.
GET /deploys/catalog/apps -> bridge /api/catalog/apps (73 embedded apps,
single source of truth). POST /deploys/catalog/deploy {app,port,env} ->
validated (service-id charset, port 1024-65535, env name/value rules) then
proxied to bridge /api/catalog/deploy which runs the shipdeck CLI catalog
deploy (OCI image -> hardened systemd service; NO Docker daemon). Audited
as deploy.catalog with ok/seconds. 11 table-driven route tests incl.
validation rejections never reaching the bridge + bridge-failure 502 shape.
App Selector submit rewired from /apps/deploy (Docker chain) to
/deploys/catalog/deploy — same modal, same fields (port, mediaPath, plex
claim token as env), no new UI screens. Full suite 3510/3510 (156 suites).
- DC-160: /catalog/:appId read app.image/ports/env/volumes which live under
app.docker.* — every detail response was empty. Now reads docker.* and
surfaces defaultPort/subdomain/healthCheck/mediaMount/claimToken/difficulty/
setupInstructions for the deploy form. List/search use the templates own
category field (old hand-rolled CATEGORY_MAP mis-categorized apps).
- Shield (routes/shield.js + 5-line app.js mount) committed to source after
weeks as prod-only drift; regression test asserts the mount exists.
- New composition-root e2e suite: boots REAL createApp() on an ephemeral
port, probes the historically-broken routes through the REAL mount table
(wizard wiring, shield, catalog detail, deploys bridge) — the class of bug
all 3 P7 dogfood findings belonged to. 5 tests.
- Full suite 3499/3499 (155 suites); table-driven DC-160 test covers all 73
image templates.
Dogfood finding #3: upsertPerson resolved existing records by id only.
The wizard upserts a bootstrap admin AFTER a provider sync has already
created a person for the same Tailscale user (no id passed), so the same
providerUserId was duplicated into two records. Resolve existing by
providerUserId when no id is given; bootstrap Admin group + protectedAt are
preserved on the merge. New regression test; full suite 3414/3414.
Dogfood finding #2: createTailscaleProvider() built with coordClient=null;
nothing in prod ever called attachCoordClient() (the wiring the provider
docblock promised never landed), so wizard credential validation always
failed not_configured even with a valid stored token. Pass a delegating
proxy that fetches a fresh coord client per call via the sanctioned
ctx.tailscaleCoord.getClient pattern (token re-read from credentialManager
each call). Full suite 3413/3413.
The /network/wizard mount in app.js omitted providers, so the wizard built
with providers={} and every mode selection returned 503 provider_unavailable
in production. Unit tests never caught it because they inject their own
provider map. Pass built tailscale + wireguard instances.
Full suite: 3413/3413 (152 suites).
Hermes
sami7777