- 站点工作台: 登记站点 dialog (built-in site definitions or custom).
- Site detail tabs: 整理视图 / 原站视图 / 快照与 Feed / 账号 (account list +
credential form).
- 来源健康 (was 来源/Feed/快照): cross-site health table + all feeds; site
names deep-link to the site's 快照与 Feed tab. Feed table extracted to
components/workbench/feed-table.tsx.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
The committed types.ts had the pi_upstream interface spliced into the middle of
PolicyRevisionTable (and upstream_id / DB entry in the wrong places), which broke
typecheck at HEAD. Regenerated with gen:types against a database migrated to 0028.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
- cards grouped by collect / download / organize / maintain / agent checks: switch,
in-place interval or cron edit, last result and duration, next-run countdown,
failure reason, run now, skip next, reset to default, history drawer
- top stats (enabled / failed / running / total) and a scheduler-offline notice;
refreshed by the schedule.changed SSE event with a 30s fallback
- agent check cards show goal template, trigger, budget and allowlist size; create /
edit dialog (read-only capabilities, budget, trigger, interval, report to inbox)
- nav entry in the system group, route and smoke route
Co-Authored-By: Claude Opus 5.5 <[email protected]>
- modules/schedules: registerScheduledJob({key, defaults, run}) registry; defaults are
synced into scheduled_job without overriding user-changed enabled/schedule
- scheduler runs only on the io worker, one instance via pg_try_advisory_lock held on a
dedicated connection; due = enabled and next_run_at reached, or run requested;
jobs run in parallel, never re-entrant; interval next = finish + period, cron next match;
run-now does not move next_run_at and merges duplicate requests; skip advances one period;
history trimmed to the last 50 runs; interrupted runs marked failed on takeover
- agent_check jobs: trigger condition checked first, goal template rendered, runAgent
enqueued via startRun(trigger=schedule) with allowlist and budget, one run per time slot;
allowRecover uses the autoInvestigate grant check; results harvested back into the run
history and optionally reported as an info decision; 3 built-in checks, off by default
- sources-identity (feeds.poll, sites.snapshot, identity.review), acquisition-organize
(downloads.sync, cache.threshold, downloads.orchestrate, slots.reevaluate,
subscriptions.search, subscriptions.metadata, library.reconcile) and file-manager
(trash.purge) register jobs instead of running their own setInterval; default
intervals unchanged; backup registered as an unconfigured placeholder
- endpoints under /schedules (admin, console session only)
TODO: pi refreshPiTokens and auth-settings session cleanup still own their timers.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
- /pi/providers, /pi/upstreams CRUD, test (auth-only by default, no cost), models
- subscription login via pi models.login with a suspended interaction: browser mode
waits for the pasted redirect URL, device mode polls server-side; per-provider
login lock, readable callback-port-in-use errors
- credential store locks the pi_upstream row (select for update) so api and worker
refreshes never double-write a rotated refresh token
- refreshPiTokens task: refresh before expiry; failure marks expired and opens a
relogin decision, closed again after a successful login
- runAgent picks upstreamId > agent_run.upstream_id > default upstream > legacy
settings; credentials injected by the server, never logged
Co-Authored-By: Claude Opus 5.5 <[email protected]>
- PiUpstreamClient: one Models per upstream holding only its provider; credential
store adapter maps to that upstream's secret, so several accounts of one provider
coexist; host env is only read for authKind none (ambient cloud creds)
- provider catalog from builtinProviders() plus static oauth modes, env fields and
subscription terms notes; openai-compatible custom endpoints
- fake upstream provider with mock OAuth (browser paste / device approval)
- PiAgentRuntime accepts an upstream; thinking level clamped to the model
Co-Authored-By: Claude Opus 5.5 <[email protected]>
- Bottom-right FAB (56px, prism ring) with active-run count, pulse and latest
status summary; turns into "collapse" beside the open drawer; hidden on the
Agent page and while the mini dock shows; lifts above bottom floating bars
- Ctrl/Cmd+J toggles the drawer, Esc closes; input is focused on open
- Emphasized Agent entry at the top of the sidebar (expanded and collapsed)
with active-run count; header button kept only below xl
- Drawer first screen offers 3 page-aware quick goals that fill the input
- Respects reduced motion; AgentShortcutsHelp exported for the settings page
Co-Authored-By: Claude Opus 5.5 <[email protected]>
The throttle could drop the final copy heartbeat when it landed inside the
interval (fsproxy.test saw 6291456 instead of 6291461 under ci load). Phase
completion (bytes >= total) is now sent immediately like phase changes.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
- POST /files/probe writes media_asset.media_info, so scope organize.read -> organize.execute
- e2e e6-w4a-endpoints -> e10-w4a-endpoints (number clashed with e6-auto-recover)
- trash: reusing a consumed approval / retired planId is rejected and the file stays
Co-Authored-By: Claude Opus 5.5 <[email protected]>
- invoke-limits: sharedInvokeLimiter() (one semaphore set per process); agent
capability routes default to it
- /files/identify: catalog.lookup with includeRemote runs under the shared
TMDB semaphore + 15s timeout; at most 20 distinct remote queries per
request, further items fall back to local recall with a warning
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Account popover (cookie sessions) gets a "二次验证" item showing the current
state; the dialog wires GET/POST /auth/totp: begin shows the one-time base32
secret (grouped) and otpauth URI for manual entry, enable confirms with a
6-digit code, disable takes current password + code. No QR image (would need
a new dependency); session query is invalidated so the state label updates.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
S-1 (/files/move, /files/organize action=move):
- organize-guard: authorizeMoveEntities checks every source's media entities
(declared entity plus entities of registered assets under the source);
one-time approval still covers the whole request; any source that cannot
be attributed to an entity only passes with an approval; denial lists every
unauthorized item; optional non-consuming pre-check
- transfer: entities per source (directory = registered assets beneath it)
- organize: all move groups pre-checked before any group executes; any
denial -> whole request 403 approval_required
S-2:
- api.ts: GET /files/download and POST /files/trash/purge marked
auth:"session"; files routes attach requireAdminSession from the registry
annotation; handler-side purge check removed, download keeps a same-rule
defense check; permanent delete uses isAdminSession
- DEV key rule written down: equals an admin-scope key for session-only
endpoints, still admin for Move/delete approval bypass (as authorizeMove)
- api-registry.test: session list updated; new reverse assertion that no
non-session endpoint answers 403 to an admin-scope bearer
S-5 (trash restore):
- restore-to requires a writable managed library root (isWritableTarget)
- in-place restore rebuilds missing parents level by level without
following symlinks (mkdirWithin), realpath-checked per level
- cross-volume restore capped by FILES_TRASH_CROSS_MAX_BYTES; cross-volume
copies (trash in / restore out) go through fsproxy
Tests: multi-entity move 403 + listing, unattributed source, organize foreign
asset 403, download/purge 403 for DEV and admin-scope keys, restore target
not writable, symlinked parent, rebuild parents, cross-volume cap.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
- contracts: isSensitiveQueryParam (whole-name match on SENSITIVE_QUERY_PARAMS
or isSensitiveKey), the single source for query credential names
- rss/redact: drop local SENSITIVE_QUERY_KEYS; param checks go through
isSensitiveQueryParam (+ caller extras); REDACTED re-exported from contracts
- browser/redact: CREDENTIAL_PARAM built from SENSITIVE_QUERY_PARAMS, JSON keys
also checked with isSensitiveKey; keeps only its own rules (path passkey,
cookie names, a11y password fields); remove stray regex statement
- redact-consistency.test: 24 credential spellings (pass_key, c_secure_pass,
refresh_token, password, uid_key, auth, ...) masked by contracts, rss and
browser alike; plain params untouched; rss extract/unredact round-trip
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Covers W4-E4/F, W5-A..E and W6-A grouped by feature; known limitations
replaced with the 19 items from docs/plan/07 §8 (stale series-completion
item removed).
Co-Authored-By: Claude Opus 5.5 <[email protected]>
- downloads.ts observe: re-read download_task FOR UPDATE inside the tx and merge
from the locked row; qB sync could overwrite imported back to importing when
verifyAndImport committed between the sync read and write (E8 flake).
- agent finalize: accept summaryPrefix so the recovery verdict is written with
the terminal status in one update (E6 read a budget_exceeded run before the
"未恢复:…" annotation landed); recover-summary step kept, now idempotent no-op.
- organize.ts: drop local slotComplete and the once-subscription completion
write; completion/history are decided only by applySlotEvent(verify_passed)
(W5-C follow-up, removes double write). Manual slot organize uses the shared
subscriptions.slotComplete.
- vitest: server split into unit (parallel) and pg (file-serial, groupOrder 1)
projects; packages/tasks file-serial; explicit worker concurrency in e2e.
- docs: W6 stabilization record in 04-v0.1-acceptance.md.
Test expectations unchanged. Full `deno task ci` x3 green: 369s / 341s / 337s.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
- /files/move|copy go through planImport's new `transfer` variant (keep names, directory
expanded to per-file plan) → organization_record + verifyAndImport/executeRecord: per-file
ledger resumable, fsproxy, cross-volume directory move = per-file copy→verify→delete then
remove empty source dirs; shared plan store + plan_stale; asset_location follows per file,
path_binding revised when a whole directory moved (finalizeTransfer); request waits up to
FILES_TRANSFER_WAIT_SEC, else executed.pending
- /files/organize per-item overrides take effect: season/episodes (domain planOrganize
overrides), contentUnitId coverage, targetDir (no binding), operation, slotId (satisfied
after verify); warnings only for inapplicable overrides
- delete/trash/restore write organization_record (action trash|delete|restore, file-level
results; migration 0026)
- GET /files/download (admin session, nfo/subtitle/image, size cap, streamed, audited)
- GET /files/records?ids|planIds for the organize page
- fs race errors map to 503/404/403/400 instead of 500
- runbook updated
Co-Authored-By: Claude Opus 5.5 <[email protected]>
- main.ts: `--role fsworker` dispatched before the rest of the module graph is
loaded (other imports became dynamic); runs fsWorkerMain from the new
`@umt/adapters/fsworker` export (~150 ms startup in the compiled binary)
- fsproxy: default worker command picked per runtime (deno CLI -> `deno run
fsworker.ts`, compiled binary -> `<execPath> --role fsworker`); stats expose
workerKind; UMT_FS_TEST_HANG_AFTER_BYTES acceptance switch for containers
- compile: --allow-run adds /app/umt-server (spawn self); start task allows deno
- organize: replaced old file moved into the file-manager trash layout
(<root>/.umt-trash/<day>/<uuid>/<rel>) and registered in trash_item, ledger
result follows; organize.transfer log carries fsproxy stats
- settings: autoInvestigateMode (investigate/recover) + max recover attempts
- deploy.md: container verification (subprocess self, TIMEOUT -> SIGKILL x3)
Co-Authored-By: Claude Opus 5.5 <[email protected]>
- subscription sheet: 立即搜索, active-search times, 追更进度 (per-season total with source/lock,
per-episode state chips, in-flight/deferred counts) and set/lock season episode total
- labels for subscription pending and candidate decision deferred
Co-Authored-By: Claude Opus 5.5 <[email protected]>
- completion by season episode_total ledger (applySlotEvent verify_passed hook overrides organize's
complete): unknown total keeps tracking and records suspected-missing; pending never completes;
once / continuous current-season completion writes subscription_history
- evaluation: missing = release units ∩ episodeRange − covered − in-flight; series slots may hold
several unit-disjoint in-flight plans; overlapping/busy releases are deferred (slot_deferral) and
re-evaluated when a plan ends (next_evaluate_at due) or on the periodic slot re-evaluation
- per-unit upgrade via slot_unit ledger; site scope (params.siteIds); site facts in evaluation context
- season packs: after submit read qB files and setFilePriority(0) for covered/in-flight episodes;
whole pack with recorded reason when undeterminable; removed downloads release their reservation
- searchSubscription task (title/aliases × enabled sites via liveSearch, snapshot + rate limiter):
on create, every searchIntervalHours (default 24h) and POST /subscriptions/{id}/search
- refreshMetadata writes season totals (only grows, floored by covered, lock respected) and turns
metadata_pending subscriptions active; plan loop refreshes tracked series every 6h
- endpoints: subscription progress, episode-total, subscription-history; e9 continuous-series e2e
Co-Authored-By: Claude Opus 5.5 <[email protected]>
- subscription: status pending (TMDB not ready: searchable/evaluable, not completable),
last_searched_at / next_search_at
- content_unit (season): episode_total ledger with source / lock / timestamp
- acquisition_plan: unit_keys (Slot × unit reservation) and file_selection; drop the
one-in-flight-per-slot unique index (checked under slot row lock per unit instead)
- new tables slot_deferral, slot_unit (per-unit quality ledger), subscription_history
- contracts: SubscriptionParams, SubscriptionHistory, SlotDeferral, SlotUnit, SubscriptionProgress,
EpisodeTotalSet, CandidateEvaluation pri/trace/unitKeys and decision deferred;
endpoints subscriptions.search/progress/setEpisodeTotal and subscription-history
- mapping: evaluation messages via describeViolation, pri/trace passthrough
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Roots + lazy dir tree, virtualized grid with status badges and asset/slot/
download links, breadcrumb, sort, glob search, hidden toggle, filter chips,
paging and truncation notice; multi-select (Shift/Ctrl/all), action bar,
context menu, drag-to-move with dryRun preview, keyboard navigation,
offline (503) and empty/error states, mobile layout with tree drawer.
Rename/move/copy/delete dialogs with dryRun previews and reference checks,
4-step organize wizard, trash view (restore/restore-to/purge, expiry).
DEV-only ?mock=1 in-memory backend; smoke route list includes /console/files.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
- adapters/fs: FsProxy + fsworker run each file op in a killable subprocess;
progress-stall watchdog (FS_HANG_TIMEOUT_MS) for copy/hash, total timeout for
quick ops; SIGTERM -> SIGKILL -> FsError('TIMEOUT'); stat/checkRoot timeouts
never read as missing
- copyWithVerify: onExisting conflict/skip/replace (hardlink backup + atomic
rename, old file to .umt-trash or deleted), sweep stale partials before write,
document why permissions/owner are not copied
- filesMatch/filesMatchSha256 throw FsError on read failure instead of
match=false; crossDeviceMove skips redundant re-hash after sha256 copy
- FsErrorCode TIMEOUT + fsErrorClass; organize executeRecord routes fs ops
through fsproxy, retries TIMEOUT/IO up to FS_TRANSFER_MAX_ATTEMPTS then opens
a decision without deleting the record; reconcile errors no longer reported
as EEXIST
- domain: subtitle .default marker (ORGANIZE_DEFAULT_SUBTITLE_LANG) and
findOtherVersions (same season/episode/part, video only) as upgrade
disposal candidates
- scripts/bench/fs-bench.ts and benchmark record in porting doc
Co-Authored-By: Claude Opus 5.5 <[email protected]>
- Port agentui (MIT, ashish200729/agentui@b2b752b) Message/MessageBubble/MessageScroller,
StreamingResponse, ToolResult/ToolApproval, TodoList, ApprovalCard, CodeBlock/FileDiff,
Citations, PromptInput into components/agentui; drop motion/shiki/Next, switch colors,
type scale, radii, durations and easing to semantic tokens, honour reduced motion
- Conversation: streaming cursor, collapsible thinking, tool cards with kind/scope/risk/
idempotency badges and audit idempotency key, prominent server-rejection state for both
whitelist rejections and unauthorized errors, compaction markers, budget meters with
over-limit state, approval cards (admin session only), two-column final report
(model said vs system verified with task/record links), follow-scroll with jump-to-bottom
- Composer: Enter send / Shift+Enter newline, stop = abort (confirmed), append,
execute/plan_only switch, "execute this plan" from finished plan_only runs
- Session list sidebar with status filter and usage; reconnect/error banners with retry
- Drawer: route-derived context (entity/site/task), minimize to a live mini dock
- Fix stream refetch race when the run switches from loading to finished
Co-Authored-By: Claude Opus 5.5 <[email protected]>
- 0024_trash: trash_item table + unique idempotency index on artifact(files_idem)
- files.ts: TrashItemRow, delete planId/planExpiresAt, organize group warnings,
scrape returns refreshMetadata taskId
- registry tests list the new table/row schema
Co-Authored-By: Claude Opus 5.5 <[email protected]>
- POST /organization-records: sourcePaths must resolve (realpath) under a
registered library_root or downloader path-mapping host path; `..` and
symlink escapes rejected (400), dryRun included. Move requires an admin
session, or an API key with an automation_grant covering the entity or a
matching one-time approval (consumed on use), else 403. Manual Move from a
managed download_task with unsatisfied H&R returns 409 and opens a
decision; never silently downgraded to Copy.
- Series slot completion: total episode count is unknown while units are
provisional, so the slot is not satisfied and the subscription keeps
tracking; coverage is registered per actual file (creating provisional
episode units), and already_in_library files count as verified presence.
- POST /engine/runs/{id}/retry (admin): re-enqueue by original task name and
raw input with idempotency key retry:<id>:<n>, audited; retry button on
the run detail page.
- E7 regular series e2e, E8 AT24 tag removal / task disappearance e2e.
- All workspace members and root set to version 0.1.0; CHANGELOG.md.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Adds packages/contracts/src/files.ts (roots/list/stat, mkdir/rename,
move/copy/organize two-phase dryRun->planId, delete with reference checks
and trash, identify/scrape/probe) and registers Api.files in api.ts.
Implementation lands in a follow-up commit (W5-A).
Co-Authored-By: Claude Opus 5.5 <[email protected]>
