The PRIVATE_ITEMS list contained the bare 'test' entry, which rsync
interprets as 'any directory named test at any depth' — so it caught
the root-level /test/ scratch directory (intended) AND every src/test/
under each module (not intended).
Pattern is already anchored to /test (root-only). This commit rsyncs
the accumulated src/test/ tree forward so opensource has the unit tests
that have been written / updated against existing src/main/ code since
the pattern regression. Going forward each per-commit sync will carry
src/test/ files along with the main change.
OAuth token save now promotes the first available chat model when no usable default exists. Default-model resolution and provider availability checks require Provider.enabled=true alongside credentials, so disabled Providers no longer return stale defaults. /models/enabled drops the single-Provider hard-code so OpenAI OAuth and other enabled chat models surface in selectors. Docker exposes the 1455 PKCE callback via MATECLAW_OAUTH_OPENAI_CALLBACK_BIND_HOST (default 127.0.0.1) and a port mapping; deployment mode stays Host-driven.
Add a parallel effectiveAllowedToolsDisplay field on the runtime status payload so the SkillMarket detail drawer can render mcp_<server>_<slug>_<hash> with the raw tool name appended in parentheses, while leaving the original prefixed list unchanged for any caller that needs the machine name. McpSkillBridge#decorateToolNameForDisplay reverses a prefixed name via the per-server cached tool list; the frontend prefers the new display field and falls back to effectiveAllowedTools when the field is absent.
Group-chat reply slot fallback:
- The platform blocks proactive sends in group chats; outbound paths
(cron summaries, async-task completions, generated image/music/3D
delivery, TTS audio) silently failed because they fell through to
the proactive-send command. New bounded LRU maps each group chat
to its most recent inbound frame id; the dispatcher prefers that
reply slot and falls through to proactive only for single chats.
- Centralised text and media dispatch through a single helper so the
group rule never has to be re-implemented per outbound path.
Upload size pre-check + auto-downgrade:
- Without client-side limits, oversized uploads streamed for ~1 minute
before the server rejected at the finish step — users saw nothing
arrive in their chat. The new decision layer mirrors the platform's
hard limits and produces three outcomes: rejected with a friendly
reason, downgraded to a generic file delivery with an inline note,
or pass-through unchanged.
- Files over 20MB reject. Images / videos over their 10MB limit
downgrade to file. Voice content that isn't AMR or exceeds 2MB
downgrades. AMR voice within 2MB stays native.
appmsg inbound parsing:
- Forwarded complex messages (document transfers, article links,
miniprogram cards) used to fall into the inbound switch's default
branch and silently drop. The new branch flattens four sub-types
into a text marker the agent reads plus any media that needs to
reach downstream tools — document forwards reuse the same magic-
byte sniff and per-conversation upload layout as native file
inbound, so extension recovery and chat-uploads serving work
identically.
- Article links produce "[链接] title\ndescription\nurl" so the agent
can summarize without round-tripping. Miniprograms surface their
title. Unknown sub-types still emit a generic marker so the agent
is never blind.
WeCom quoted-message context:
- Parse the body.quote field that arrives alongside any inbound message
(text / image / voice / file / mixed sub-types). When a user long-
presses a previous bot bubble and types a follow-up like "解释一下",
the agent now sees both the user's new text and the referenced
content as proper context — replies stay on topic instead of
guessing what was being explained.
- Quoted images / files are downloaded through the same pipeline as
inbound new media (magic-byte sniff, ZIP container peek for
DOCX / XLSX / PPTX recovery, chat-uploads layout) so the vision
sidecar and document tools can actually analyse what was quoted.
- Reading order in the assembled prompt: "[引用消息: ...]\n<user text>"
first, then quoted media parts, then the user's own current-message
media. Mixed quotes flatten into a space-joined summary.
Multimodal sidecar settings preservation:
- The bulk settings PUT used to unconditionally overwrite the vision /
video sidecar model ids — null in a partial payload became "" in
the DB, silently wiping the configured sidecar every time a user
saved an unrelated settings page (System / Music / Image / etc.).
Symptom: "I picked a vision model, saved a different settings tab,
now the bot can't see images anymore."
- Bulk save now guards both keys with non-null checks, matching the
pattern used for music / 3D / image / video / tts / stt blocks.
- A dedicated /settings/sidecar endpoint always writes both keys, so
the sidecar UI can still explicitly clear via null without leaking
the write-on-null semantics into every other settings save.
- Frontend sidecar card switches to the dedicated endpoint; other
settings pages keep their existing partial-payload behaviour.
Inbound (WeCom):
- Save uploaded media under data/chat-uploads/{conversationId}/ with full
fileName/path/fileUrl/storedName/fileSize on the content part. Web mirrors
of an IM conversation now show real thumbnails instead of "未命名".
- Magic-byte sniff (PDF / PNG / JPEG / GIF / Office / ODF / archives /
audio / video) recovers a real extension when the platform omits filename
for forwarded files — no more PDFs labelled "file.bin".
- ZIP container peek distinguishes DOCX / XLSX / PPTX / VSDX / ODT / ODS /
ODP / EPUB / JAR from a plain zip via discriminator paths and the OASIS
mimetype entry.
Outbound (WeCom):
- Chunk upload field name corrected so server-side actually stores the
bytes — file messages used to arrive with correct filename/size but
empty content, breaking every PDF / DOCX / PPTX recipient.
- Scan agent text for served-file URLs in both the text-reply and
content-parts paths; fetch bytes from the in-memory generated-file
cache and dispatch through the native chunk upload + media message
protocol so users receive a tappable file card instead of an
unopenable markdown link. Cache miss surfaces a clear retry hint.
Async tool result forwarding:
- New AsyncTaskMediaDispatcher routes generation completions (image,
video, music, 3D model) to whichever IM channel the conversation is
bound to via ChannelSessionStore + ChannelManager. Web / webchat
conversations are intentionally skipped — their SSE stream already
renders the result.
- Wired into all four generation services so IM users actually receive
generated media as native attachments. Each part now carries an
absolute disk path so adapters read bytes locally instead of round-
tripping through an authenticated served URL.
Slack native file upload:
- SlackChannelAdapter overrides the content-parts dispatch. Image /
audio / video / file / model3d parts ride filesUploadV2 so users see a
file card with preview thumbnail bound to the same thread as the
originating message. Text parts continue through chat.postMessage.
- Resolves bytes from the part's local path, falls back to an HTTP fetch
of fully-qualified URLs.
IM approval hint visibility:
- IM-driven approve / deny / auto-cancel / replay-error hints now go
through saveMessage + tracker broadcast in addition to the channel
adapter, so a Web mirror viewing the same conversationId sees the
resolution. Previously hints reached only the IM channel; the Web
admin console had no record of the outcome.
Adaptive paste-merge debounce:
- WeCom and other IM clients silently split long pasted prompts into
fragments that arrive 0.5-2 seconds apart, missing the existing 500ms
merge window. The agent then saw torn context and emitted multiple
conflicting replies.
- When the merged buffer crosses a content-length threshold, extend the
debounce window so subsequent fragments arrive in time. Default
500ms unchanged for normal short messages.
Resolve ${user.home} and other JVM system properties in MCP server env, headers, and cwd — previously only OS env vars were expanded, causing the filesystem MCP server to fail on Windows where $HOME isn't set.
Temporarily restore McpClientManager.java to its pre-#60 state so the
contributor's PR can squash-merge cleanly with their authorship preserved.
The args-expansion follow-up will land as a separate commit right after.
When an agent had any skill bound, the runtime tool gate was silently
hiding @Tool beans that aren't declared in any skill manifest, even
though the global system prompts (SOUL.md / "Web Search Capability" /
"File Reading Guidelines") explicitly tell the LLM these tools are
available. Result: the model would call search / renderDocx / read_file
/ etc., hit "Tool not found", then either give up or fall back to
unhelpful behaviour (e.g. dumping markdown text instead of producing a
.docx download).
This commit:
- Adds universally-promised, agent-wide tools to SYSTEM_LEVEL_TOOLS so
they bypass the manifest restriction: document/media generation
(renderDocx*, image_generate, music_generate, video_generate),
global capability tools the system prompt mentions (search,
browser_use, read_file / write_file / edit_file /
execute_shell_command, detect_file_type, extract_*_text,
readMateClawDoc), skill discovery siblings (listSkillFiles,
listAvailableSkills), and the delegate triplet (delegateToAgent,
delegateParallel, listAvailableAgents).
- Fixes 5 entries in the prior whitelist whose names did not match
any real @Tool bean and were therefore silently dead:
read_workspace_file -> read_workspace_memory_file
write_workspace_file -> write_workspace_memory_file
list_workspace_files -> list_workspace_memory_files
delegate_agent -> delegateToAgent
datetime -> getCurrentDate / getCurrentDateTime / getCurrentTime
Also adds the missing edit_workspace_memory_file.
- In the chat markdown renderer, strips any hallucinated
https?://<host> prefix from /api/v1/files/generated/<id> download
links before building the <a href>. Multiple LLMs have been
observed prepending bogus hosts when echoing tool-returned download
URLs back to the user, breaking the click. One-line defensive
normalization independent of which model is in use.
Verified end-to-end on a previously-broken agent: search / browser_use
/ execute_shell_command / renderDocx all dispatch correctly now and
the final markdown link is a clean same-origin path. 36 whitelist
entries cross-checked against real @Tool method names.
AgentBindingServiceTest green.
When a user-installed skill (e.g. RedisOps) was bound to an agent, the
model frequently called the skill name directly as a tool, hit
"Tool not found: RedisOps", and either gave up or fell back to shell
guessing. Two compounding causes:
1. The system prompt block injected by SkillRuntimeService listed each
skill as `- **RedisOps** — desc`, which is the same format used for
tool catalogs and primed the model to call the names directly. The
"how to use" instructions referenced `read_skill_file` /
`run_skill_script` — names that don't exist in the tool registry,
so even a compliant LLM couldn't follow them.
2. ToolExecutionExecutor's `callback == null` branches returned a bare
"Tool not found: <name>" string. The model had no recovery signal
and no hint that the name it called was actually a skill.
Fix is two-layered:
- Prompt rewrite (SkillRuntimeService.buildSkillPromptEnhancement): lead
with an explicit warning that skills are NOT directly callable, use the
correct camelCase tool names (readSkillFile / runSkillScript), include
a concrete worked example anchored to the first enabled skill, and
render the listing as a markdown table so it stops looking like a
callable tool list. listAvailableSkills tool description and output
follow the same pattern.
- Runtime safety net (ToolExecutionExecutor): when toolCallbackMap.get
misses, check if the requested name (case-insensitive) matches an
active skill. If so, return a precise hint telling the LLM the right
invocation pattern instead of the bare error. Wired through both the
main execute path and the pre-approved replay path. SkillRuntimeService
is attached via a setter from AgentGraphBuilder so the executor's many
legacy constructors stay untouched, and it's nullable so isolated
tests still work.
Adds 5 unit tests covering: skill match -> hint, case-insensitive match,
no-match -> bare error, no SkillRuntimeService wired -> bare error,
pre-approved replay path -> hint.
Reported and reproduced by @pipima9950-glitch in issue #46.
SkillPackageResolver.persistScanOutcome built a fresh SkillEntity with only
id + scan fields, then called updateById. SkillEntity declares six columns
with @TableField(updateStrategy = FieldStrategy.ALWAYS) — name_zh, name_en,
config_json, source_code, skill_content, security_scan_result — so the
ALWAYS strategy emits UPDATE statements that write NULL to every one of
those columns not set on the partial entity.
Effect: every security re-scan that produced a status/findings change
silently wiped skill_content, config_json, source_code, name_zh, name_en
on the row. After importing a custom skill, the first scan tick destroyed
the imported content.
Fix: switch to LambdaUpdateWrapper so the UPDATE only touches the three
scan columns we actually want to change. Other skillMapper.updateById
call sites (SkillService, BuiltinSkillSeedService) pass DB-hydrated
existing entities and are unaffected.
Reported and diagnosed by @pipima9950-glitch in issue #45.
Three layers landed together because they share the same routing /
lifecycle plumbing:
1. Cron output unification
- New CronConversationResolver routes web-origin jobs to the per-workspace
tasks_<wsId> conversation; IM-bound jobs go to the channel session
conversation when one exists (matched by senderId then targetId);
legacy cron_<id> remains as the fallback.
- CronJobLifecycleService inserts a system-role header divider when a
run starts so users browsing the unified tasks_<wsId> view can tell
which job started a run. BaseAgent.sanitizeForLlm filters these
headers so they never reach the model.
- WorkspaceService seeds tasks_<wsId> on workspace creation; V65
migration backfills existing workspaces.
- DeliveryConfig gains a userId field so IM session lookup can match
by senderId (replyToken-based targetId is not stable across runs).
- ConversationVO recognizes tasks_/cron_ underscore prefix as cron
source. MessageList renders the system header as a labeled divider.
- ChatConsole pins tasks_* conversations and tracks per-conversation
read state so new cron output gets a visible unread dot.
2. Reminder task type
- New task_type='reminder' in CronJobEntity + service validation.
- CronJobRunner short-circuits 'reminder' jobs: hands trigger_message
to finishRunAndPublish verbatim, no LLM call. Fixes a regression
where reminders were rephrased into echoed wrappers.
- New create_reminder tool alongside create_cron_job, with descriptions
tightened so the model picks the right one (verbatim push vs LLM
query that needs computation).
- CronJobs.vue gets a third radio option + dedicated reminder field.
3. In-flight progress placeholder
- Cron uses non-streaming chat()/execute(); tool-heavy ReAct loops
can run 1-5 minutes between start and finish with no visible
state, looking hung.
- New GET /api/v1/cron-jobs/active-runs returns runs in status=running
for a conversation. ChatConsole polls it on the existing 4s tick
(and on conversation switch) and shows a spinner bar with elapsed
time. When run count drops to zero, it refetches messages so the
assistant bubble appears within ~1s of finish.
Add a callout above the existing intro to make the wedge explicit:
multi-user workspaces, approval-gated sensitive actions, full audit trail,
production-grade health monitoring, per-channel error isolation.
One JAR on your own machine, zero data egress.
Chat attachments with non-ASCII filenames (e.g. Chinese) get sanitized
at upload time — `人人有虾.docx` is stored as `1777391026594_____.docx`.
Tools then receive only the original filename via '[Attachment] foo.docx'
and fail with 'file not found'.
- renderMessageContent now appends the actual server-side path so any
tool the LLM picks (read_file / extract_document_text /
detect_file_type) gets a path that resolves directly.
- New ChatUploadResolver helper performs basename-suffix matching inside
the conversation's chat-upload directory; ReadFileTool, DocumentExtractTool
and FileTypeDetectorTool fall through to it when the literal path does
not exist (defense in depth for cases where the LLM ignores the path
hint).
Refs https://github.com/matevip/mateclaw/issues/29
Replace the per-startup admin reconciliation in
WorkspaceSchemaMigration.ensureDefaultWorkspaceMembership() with a
one-shot bootstrap. Once the default workspace has any owner, the
method returns immediately, so an operator's deliberate removal of an
admin from the default workspace persists across restarts. If no owner
exists yet, pick the lowest-id active admin and add them as owner; if
no admin exists at all, log a warning and skip rather than failing
startup.
Refs https://github.com/matevip/mateclaw/issues/29
Restart-time backfill in WorkspaceSchemaMigration was inserting every
existing user into the default workspace and copying mate_user.role
('user'/'admin') into mate_workspace_member.role, whose valid domain is
{owner, admin, member, viewer}. Result: non-admin users assigned to
other workspaces were silently re-attached to the default workspace
with role='user', failing roleLevel() lookup and 403'ing on Agents.
- Filter the INSERT on u.role = 'admin' and hard-code the membership
role to 'owner', removing the role-domain mismatch and the
workspace-isolation violation in one change.
- Add V60__fix_invalid_workspace_member_roles.sql (h2 + mysql) to
drop already-corrupted default-workspace rows for users who have a
valid membership elsewhere, and downgrade the orphan rows to
'member' so those users aren't locked out entirely.
Refs https://github.com/matevip/mateclaw/issues/29
User-reported field issues + a deeper code audit revealed multiple
overlapping bugs in the prior cron-channel delivery change. This fixes
all six.
#1 — Concurrency race on ToolExecutionExecutor (root cause of 'sometimes
succeeds, sometimes fails' tool calls). The volatile instance fields
currentRequesterId / currentWorkspaceBasePath / currentChatOrigin
were shared by every conversation routed through the same per-agent
executor; one user mid-build-loop while another's execute()
overwrote the field would cross-contaminate the captured values into
PreparedToolCall. Fix: kill the instance fields, thread
origin/requester/workspace as method params straight into
PreparedToolCall snapshot. Comment pins the rule so it cannot regress.
#2 — CHAT_ORIGIN missing from KeyStrategyFactory (latent timebomb,
masked by spring-ai-alibaba-graph-core's non-filtering builder path).
Without an addStrategy registration, multi-node state merges in long
ReAct / Plan-Execute loops drop the key, ActionNode reads
ChatOrigin.EMPTY, and the cron persists with channel_id=NULL. Also
caught 4 more keys that were latently unregistered:
WORKSPACE_BASE_PATH, STOP_REQUESTED, RETURN_DIRECT_TRIGGERED,
DIRECT_TOOL_OUTPUTS. All five now registered in both ReAct and
Plan-Execute factories.
#3 — CronJobs UI didn't surface channel binding. CronJobDTO carried
channelId / deliveryConfig but the list page never rendered them.
Added: (a) 'channel' column on list page, (b) channel + targetId
rows in the detail modal, (c) backend batch-loads channel names via
ChannelMapper.selectBatchIds so the column shows the human-readable
name, (d) i18n keys (zh + en), (e) channelName field on TS CronJob
type.
#4a — DingTalk targetId expiry. ChannelChatOriginFactory.resolveTargetId
used to prefer ChannelMessage.replyToken which for DingTalk encodes
a sessionWebhook URL that expires ~90 minutes after the inbound
message. Cron persisted with that webhook then dies with 401/403 and
marks NOT_DELIVERED forever. Fix: prefer the stable chatId, fall
back to senderId — both work indefinitely via DingTalk's Robot API.
#4b — Scheduler pool exhaustion under long LLM. CronJobService's
ThreadPoolTaskScheduler ran with poolSize=4 AND the LLM call lived
on the scheduler thread. Four concurrent crons saturated the pool
and the 5th silently missed its tick. Fix: keep scheduler tiny (it
just fires triggers) and offload runAgent to a dedicated
virtual-thread executor (cron-execute-* threads). LLM workload is
I/O-bound — virtual threads scale to thousands at trivial cost.
#5 — Minor latent bugs:
- AbstractCronResultDelivery.claimRun used .in(... 'NONE','PENDING',null),
but SQL IN never matches NULL. Rewrote as IS NULL OR IN
(NONE,PENDING) so legacy pre-V57 rows can still claim.
- CronDeliveryListener.onCompletedRaw was an empty @EventListener
with a wrong-headed comment about test fallbackExecution. Removed.
- CronJobTool.resolveAgentId silently returned 1L when origin
lacked an agentId — would silently bind to whatever agent #1
happens to be. Replaced with explicit error so wiring bugs surface
immediately instead of producing scheduled-but-never-runs crons.
State-key registration guard. New StateKeyRegistrationCoverageTest
scans MateClawStateKeys via reflection and parses
AgentGraphBuilder.java to extract every
.addStrategy(MateClawStateKeys.X, ...). Asserts every non-_NODE
constant appears in at least one factory. Caught the 4 unregistered
keys above on first run; will catch any future 'forgot to register'
regression.
Tests: 33 unit/arch tests + 27 regression in touched areas — all green.
Vue typecheck clean.
Refs: #25, #16
Replaces the prior ThreadLocal context plumbing with explicit Spring AI
ToolContext threading carried by an immutable ChatOrigin value object,
so a cron created from inside WeChat (or any IM channel) delivers its
results back to the originating channel.
Architecture
- ChatOrigin / ChannelTarget value objects + per-entry-point factories
(ChannelChatOriginFactory in vip.mate.channel, CronChatOriginFactory
in vip.mate.cron — symmetric, no cyclic deps).
- LocaleAwareToolCallback now forwards call(String, ToolContext) and
getToolMetadata so the decorator chain cannot silently drop the origin.
- AgentService 6-method overhaul + ChatOriginHolder bridge into
StateGraph buildInitialState which writes CHAT_ORIGIN; ActionNode +
StepExecutionNode forward it to ToolExecutionExecutor.
- ToolExecutionExecutor builds ToolContext per call; 8/8 tools migrated
(CronJobTool, WorkspacePathGuard, Video/Image/Browser/ReadFile/Music,
DelegateAgentTool with parent-origin inheritance).
- CronJobRunner + CronJobLifecycleService 3-segment REQUIRES_NEW model
(T1 startRun / no-tx runAgent / T2 finishRunAndPublish); ArchUnit
pins CronJobRunner as @Transactional-free.
- CronResultDelivery Strategy + AbstractCronResultDelivery Template
with SQL CAS idempotency on mate_cron_job_run.delivery_status —
replaces the prior process-local Caffeine TTL, cluster-safe.
- CronJobCompletedEvent + @Async @TransactionalEventListener(AFTER_COMMIT);
cronDeliveryExecutor (core=2, max=4, queue=1000, AbortPolicy + audit).
- CronRunStaleCleanup @Scheduled(5min) sweeps PENDING-15min and
status='running'-30min in one query each.
- CronJobRunner.wrapWithDeliveryGuard prepends a system note for
channel-bound crons to suppress hallucinated 'install CLI to send
WeChat' suggestions.
- ApprovalWorkflowService Memento: persist ChatOrigin snapshot on
create, restore on replay so cross-restart approvals keep channel
binding; ChannelMessageRouter + ChatController web-replay both prefer
the Memento and fall back to fresh-build.
- ChannelManager.sendToChannel 4-arg DeliveryOptions overload;
ChannelAdapter#proactiveSend default 4-arg pass-through; Slack
overrides for thread_ts and Telegram overrides for message_thread_id.
- CronJobs UI: read-only 'last delivery' badge driven by
CronJobMapper.selectListWithDeliveryStatus subquery.
Schema migrations V57/V58/V59 (V56 was already taken by an unrelated
provider migration — Flyway processes versions in order regardless of
gaps):
- V57: mate_cron_job_run delivery_status / target / error + composite
index (delivery_status, started_at) covering the cleanup sweep.
- V58: mate_cron_job channel_id (indexed) + delivery_config TEXT (JSON
via MyBatis Plus JacksonTypeHandler).
- V59: mate_tool_approval chat_origin TEXT (Memento).
All idempotent in both H2 (IF NOT EXISTS) and MySQL (INFORMATION_SCHEMA
guard + PREPARE).
ArchUnit guards (test scope, archunit-junit5 1.3.0):
- every concrete vip.mate.* ToolCallback must override
call(String, ToolContext) — pins the decorator-forward fix.
- CronJobRunner must NOT carry @Transactional on the class or any
method — pins the 3-segment lifecycle rule.
Tests: 32 new unit tests + 21 regression tests in touched areas, all
53 green:
- ChatOriginTest (6) — value-object invariants + JSON round-trip.
- LocaleAwareToolCallbackToolContextTest (2) — decorator forward.
- DeliveryConfigTest (4) — Jackson round-trip + forward-compat.
- ToolCallbackToolContextForwardArchTest (2) — both ArchUnit guards.
- CronJobRunnerDeliveryGuardTest (3) — channel-cron prefix injection.
- AbstractCronResultDeliveryTest (4) — claim CAS + concurrent CAS.
- ChannelCronResultDeliveryTest (6) — supports / doDeliver / errors.
- ApprovalReplayContinuityTest (5) — Memento round-trip + corrupt
payload fallback + unknown-field tolerance.
Refs: #25, #16
Volcano Ark exposes a separate 'Coding Plan' subscription endpoint at
/api/coding/v3 with its own coding-tuned model catalog (ark-code-latest,
doubao-seed-code, kimi-k2-thinking, glm-4.7 coding edition, etc.). The
same Volcano API key works against it. Splitting into a sibling
volcengine-plan provider lets users keep chat-tuned and coding-tuned
defaults side by side, and the generalized OpenAI-compatible path
resolver already handles the /v3 suffix without a completionsPath
override.
Adds Flyway V56 (h2 + mysql) and updates the 4 seed-data files with
matching rows (ids 1000000320-325) for fresh installs.
- Generalize the OpenAI-compatible chat/models path resolver so any
baseUrl ending in /v{N} (Ark /v3, Zhipu /v4, ...) drops the duplicate
/v1 prefix. Volcano Engine test-connection and chat were posting to
/api/v3/v1/chat/completions and getting 404.
- Replace the six pre-seeded Doubao alias rows (doubao-1.5-*) with five
valid Ark direct-call ids (doubao-seed-1-8-251228 etc.) and flip
support_model_discovery=TRUE so users can refresh their account's
actual catalog. Aliases were marketing names, not API names, so every
call hit InvalidEndpointOrModel.NotFound.
- Translate Ark business errors into actionable Chinese hints: include
the response body in the error chain, match ModelNotOpen and
InvalidEndpointOrModel codes, extract the offending model id, and
classify them as MODEL_NOT_FOUND so failover skips retries.
Adds explicit user-intent gating to the provider catalog. Fresh installs
get an empty dropdown by default — built-in cloud + local providers
(OpenAI, Anthropic, Ollama, LM Studio, MLX, llama.cpp, etc.) live in a
new 'Add Provider' drawer until the user opts them in. Existing installs
upgrade conservatively: V55 promotes any provider with evidence of use
(real api_key, OAuth token, recent chat usage, or current default model).
Backend
- V55 migration (H2 + MySQL): adds enabled BOOLEAN DEFAULT FALSE on
mate_model_provider, plus 4 promote-to-true UPDATE rules. Also
CREATE INDEX idx_message_runtime_provider_time so the 30-day usage
lookup doesn't full-scan mate_message on heavy users.
- ModelProviderEntity, ProviderInfoDTO: enabled field.
- ModelProviderService:
* listProviders() now filters WHERE enabled = TRUE — chat path,
ModelSelector, Settings/Models main grid see only opted-in rows.
* listCatalog() new — full catalog (enabled + disabled) for the drawer.
* setEnabled(id, enabled) flips the flag, publishes
ModelConfigChangedEvent (re-probe via the existing listener), and
on disable auto-promotes a replacement default model when the
disabled provider owned the current default. Returns EnableResult
so the frontend can fire a toast.
* createCustomProvider sets enabled=true (user just made the row).
- ProviderInitProbe.listConfiguredProviders also filters enabled=true —
no point probing rows the user can't see.
- ModelConfigController: GET /catalog, POST /{id}/enable, POST /{id}/disable.
- Plugin-registered ChatModels are unaffected — they live in
pluginChatModels (in-memory map), don't go through DB listProviders,
so the enabled filter doesn't strand them.
Frontend
- New types: ProviderInfo.enabled, EnableResult.
- New API: catalog / enableProvider / disableProvider.
- New composable useProviderEnablement: catalog ref, drawerOpen,
togglingId, loadCatalog, openDrawer / closeDrawer, enableProvider,
disableProvider (fires defaultSwitchedToast on auto-switch).
- AddProviderDrawer.vue: lazy-loaded, reuses DoctorDrawer's Teleport +
overlay + slide-in panel pattern. Two groups (cloud / local),
unenabled rows surface to the top of each group, enabled rows show
an 'Enabled' badge instead of a button. Mobile: full-screen sheet
that slides up from below.
- ProviderCard: new 'Disable' button with soft-danger styling on
enabled providers — soft-hide that keeps the config; user can
re-enable from the drawer.
- Settings/Models index.vue:
* Two top CTAs: 'Enable Provider' (drawer) and 'Custom' (existing
custom-create modal) — distinct workflows, both surfaced.
* Empty state with prominent 'Enable Provider' CTA when zero
enabled providers — paired with onMounted auto-open of the
drawer (sessionStorage guard so closing it doesn't bring it
back on the next route visit in the same session).
* Deep-link: ?addProvider=1 query forces the drawer open and
strips itself after, so a back/forward doesn't re-fire the open.
- ModelSelector: when groups.length === 0 and not searching, show
'No providers configured -> Configure' CTA linking to
/settings/models?addProvider=1 — the natural flow when a fresh
user opens chat before configuring anything.
- i18n: 13 new keys per locale (zh-CN + en-US) plus common.close.
Migration safety
- Conservative default policy: only rows with concrete evidence of
use are auto-enabled; everything else stays hidden. Upgrade users
may notice unused built-ins disappearing from their dropdown —
that's the intended cleanup.
- mate_message index added so the 30-day usage rule doesn't full-scan
on large installations; FlywayRepairConfig handles redeploy idempotency.
Tests
- ModelProviderServiceEnableTest covers all 7 enable/disable branches:
flag flip + event publish, no-op on already-{enabled,disabled},
default-switch when disabled provider owned current default,
no-switch when default belongs elsewhere, no-replacement returns
unchanged, getDefaultModel exception path, candidates with no
models are skipped.
- ProviderInitProbeTest: helper provider() now sets enabled=true so
the new probe filter doesn't strand existing fixtures.
- vip.mate.llm.** suite: 125 tests green. vue-tsc 0 errors. Browser
page renders with both new buttons + drawer.
Three knots untangled so an image sent from DingTalk lands in both the
LLM's multimodal prompt and the chat history bubble:
- Prefer MessageContent.downloadCode (universal, used by the new
api.dingtalk.com messageFiles/download) over pictureDownloadCode
(legacy oapi field). Sending the legacy code to the new API got
HTTP 500 unknownError, which was the original 'image not recognized'.
- After fetching bytes, persist to ~/.mateclaw/media/dingtalk/ so vision
can read via FileSystemResource, AND stuff the same bytes into
GeneratedFileCache so the UI gets an /api/v1/files/generated/{id} URL
to render. Without the URL the message bubble showed an empty card.
- Carry filename / contentType / size on the MessageContentPart so the
chat history doesn't fall back to the 'unknown' caption.
Same treatment applied to the richText branch (inline images from the
PC client) and threaded through the Stream SDK path.
Bundles in the prerequisite ChannelManager wiring of GeneratedFileCache
into DingTalkChannelAdapter and the new DingTalkMediaUploader used by
the outbound attachment flow that this work depends on.
Known limit: GeneratedFileCache TTL is 10 min — fresh refreshes work,
but viewing the image after a JVM restart needs a stable on-disk
serving endpoint, which is intentionally out of scope here.
Phase 1 of the model-module refactor: combine pool / cooldown / probe-
completion signals into a single Liveness state surfaced through the
provider DTO, so the dropdown stops listing providers that are provably
unreachable. Zero schema change; one PR backend + frontend.
Backend
- Liveness enum with five mutually-exclusive states: LIVE, COOLDOWN,
REMOVED, UNPROBED, UNCONFIGURED. Computed in ModelProviderService
from AvailableProviderPool / ProviderHealthTracker / ProviderInitProbe
snapshots batched once per listProviders() call.
- ProviderInitProbe.hasBeenProbed exposes a monotonic Set so the UI
can distinguish 'still booting' from 'probed and removed' — without
it the startup window flashes false REMOVED states.
- ProviderInfoDTO gains liveness + unavailableReason +
cooldownRemainingMs + lastProbedAtMs. The legacy 'available' boolean
stays but is now derived from liveness == LIVE so the chat fallback
walker and the dropdown agree about what's usable.
- ProviderInitProbe injected into ModelProviderService via
ObjectProvider to break the startup cycle (probe already depends on
the service).
Frontend
- ProviderInfo type extended with liveness + the three detail fields.
- ModelSelector filters UNCONFIGURED + REMOVED out of the dropdown,
shows COOLDOWN / UNPROBED with a status dot and dimmed rows that the
user can still click to override.
- ProviderCard renders a five-state badge driven by liveness instead
of the old configured + pool-entry combo. Reprobe button now keys
off liveness in {REMOVED, COOLDOWN}.
- useProviders drops loadProviderPool / providerPool — pool data ships
inline on each ProviderInfo, saves a round trip per page load and
keeps a single source of truth.
- i18n: 8 new keys across zh-CN and en-US for liveness labels and the
cooldown countdown tooltips.
Bonus fix (discovered during verification): AgentGraphBuilder.buildOpenAiApi
hard-required a usable API key on every OpenAI-compat provider, ignoring
the per-provider requireApiKey flag. That bug stranded keyless local
runtimes (LM Studio / MLX / llama.cpp) the moment a user actually
launched them; Ollama only worked by accident because its seed row
carries a placeholder string in api_key. keyRequired now honors
requireApiKey, and Spring AI's NoopApiKey is used when no key is needed
so the Authorization header is omitted entirely.
Test
- ModelProviderServiceLivenessTest covers all five Liveness states +
the probe-bean-absent fallback branch.
- vip.mate.llm.** suite (118 tests) green; vue-tsc clean.
- End-to-end browser sanity: 27 raw providers reduce to 6 LIVE groups
in the chat dropdown; LM Studio / MLX / llama.cpp render REMOVED red
badges with reprobe buttons; cloud providers without keys show
UNCONFIGURED.
The stream SDK delivers voice messages as ChatbotMessage with msgtype=audio
and the server-side ASR result already filled into MessageContent.recognition
(same shape as WeCom's voice.content). The adapter's handleStreamMessage
only read msg.getText(), which is null for audio events, so the message
landed in handleWebhook with no msgtype, fell through to the default text
branch, found null content, and got dropped at 'Empty message content,
ignoring'. From the user's side: send a voice, nothing happens, no log of
the attempt.
Two surgical edits:
- handleStreamMessage now checks getContent().getRecognition() first; if
present and non-blank, builds payload {msgtype: audio, audio: {recognition}}
before falling back to the existing text path. The earlier comment about
richText being handled inside handleWebhook was wrong — picture and
richText also need their fields propagated through the payload Map; left
a TODO for them.
- handleWebhook gains an explicit case 'audio' branch that pulls text out
of audio.recognition and pushes it onto contentParts.
- ChannelMessage.inputMode now reflects 'voice' when msgtype=audio,
mirroring feishu's behavior so downstream code (memory-extraction
filters, voice-themed system prompts) can tell text vs voice turns apart.
No STT call required — DingTalk transcribes server-side and ships text in
the webhook, so this is a 0-network, 0-config fix.
Mirrors the feishu one-click flow: scan a QR with the DingTalk app,
approve, and the bot's client_id / client_secret get auto-filled instead
of forcing the user through the open-dev console. Saves about seven
manual steps per channel setup.
Backend
- Bump dingtalk-stream from 1.3.5 to 1.3.12. Diff against the classes we
depend on (OpenDingTalkStreamClient, ChatbotMessage, MessageContent,
GenericEventListener) is empty — pure point-release bumps, no API churn.
- New DingTalkAppRegistrationService: synchronously runs init + begin
against /app/registration/{init,begin} on oapi.dingtalk.com to obtain
the device_code and verification URL, then spawns a daemon worker that
polls /app/registration/poll every 5s until SUCCESS / FAIL / EXPIRED is
returned. Sessions evict after 7 minutes, worker has a 6-minute hard
runtime cap, transient HTTP errors do not terminate the loop. Same
shape as the feishu service, but written from scratch because the
dingtalk-stream SDK doesn't wrap this OAuth device flow.
- Two new endpoints under /api/v1/channels/webhook:
POST /dingtalk/register/begin returns session_id;
GET /dingtalk/register/status returns status + qrcode_img (data URI
PNG, ZXing-encoded from the verification URL, matching the feishu and
weixin flows). Status surface: waiting / confirmed / expired / denied.
Frontend
- channelApi.dingtalkRegisterBegin / dingtalkRegisterStatus.
- New useDingTalkAppRegister composable, structurally identical to
useFeishuAppRegister minus the domain argument. Stops polling on
terminal status, fires onConfirmed with {clientId, clientSecret}.
- ChannelEditModal: dingtalk-register-card rendered when channelType is
dingtalk, scoped DingTalk blue (#1f79ff) to differentiate from feishu's
indigo. onConfirmed writes channelConfig.client_id / client_secret so
the existing form fields update reactively.
- i18n: channels.dingtalkRegister.* keys for title / hint / button states
/ scan / confirmed / expired / denied / startFailed.
Saves the user the entire 'go to the open platform -> create an enterprise
app -> copy App ID and Secret' detour. Click a button in the channel form,
scan the QR code, confirm authorization, credentials are auto-filled.
Backend
- Bump com.larksuite.oapi:oapi-sdk from 2.5.3 to 2.6.1, which adds the
scene/registration package wrapping the device-code flow.
- New FeishuAppRegistrationService: each begin() creates a sessionId,
spawns a worker thread, runs the SDK's blocking RegisterApp.register
with onQRCode and onStatusChange wired into a per-session state machine
(PENDING -> WAITING -> CONFIRMED / EXPIRED / DENIED / ERROR). The
session caches the QR data URI so ZXing only encodes once per attempt.
Sessions evict after 5 minutes so closed browsers don't leak the map.
- Two new webhook endpoints under /api/v1/channels/webhook/feishu:
POST /register/begin returns session_id, GET /register/status returns
status + qrcode_img (data URI base64 PNG, ZXing-encoded from the SDK's
verification URL — the raw URL would render as a broken image, so the
encoding step matches the WeCom flow).
- SDK detail caught the hard way: don't pass .domain() or .larkDomain().
The SDK defaults are accounts.feishu.cn / accounts.larksuite.com (the
registration endpoints). open.feishu.cn is the open-API endpoint, a
completely different service. Passing the wrong one makes the SDK parse
HTML as JSON and emit invalid_response.
Frontend
- channelApi: feishuRegisterBegin / feishuRegisterStatus.
- New useFeishuAppRegister composable: state machine that begins the
session, polls status every 2s, prefers qrcode_img over qrcode_url for
the <img> src, stops on terminal status, fires onConfirmed with
{appId, appSecret}.
- ChannelEditModal: a new feishu-register-card above the wecom one. The
composable's onConfirmed writes channelConfig.app_id / app_secret, so
the existing form fields update reactively.
- i18n: channels.feishuRegister.* keys for title / hint / button states /
scan / confirmed / expired / denied / error.
The zh seed planted channels with English display names (DingTalk Bot,
Feishu Bot, WeCom Bot, ...). The type label localized correctly but the
per-channel name stayed English on the cards page even when UI was Chinese.
- Update zh seed files (data-zh.sql + data-mysql-zh.sql) so fresh installs
get Chinese names from the start: Web 控制台, 钉钉机器人, 飞书机器人,
Telegram 机器人, Discord 机器人, 企业微信机器人, QQ 机器人, Slack 机器人.
id=1000000008 (微信) was already Chinese; left alone. en seeds untouched.
- Add V54 migration that flips existing zh-CN installs in place. Each
UPDATE is gated on system_setting language=zh-CN AND the channel name
still equal to its original English seeded value, so user-renamed
channels are left alone. Subsequent runs match no rows (idempotent).
h2 and mysql variants stay in lockstep.
Backend (FeishuChannelAdapter):
- Default connection_mode flips webhook -> websocket on doStart and doReconnect.
- Stale event filter: drop events whose message.create_time is older than
stale_event_threshold_seconds (default 30s) so SDK reconnect replays do not
re-trigger the agent.
- Silent disconnect watchdog runs every 60s; if no events arrive for
silent_disconnect_threshold_seconds (default 1800s) after the first event,
call onDisconnected to force a reconnect cycle. Setting the threshold to 0
disables the watchdog. The watchdog is scheduled before wsClient.start() on
the bring-up path because that call blocks indefinitely.
- Quoted message context: when a reply has parent_id set, fetch the parent
via GET /open-apis/im/v1/messages/{id}, summarize per msg_type (text / post
first paragraph / [Image]/[File]/[Audio]/[Video] placeholders, capped at
200 chars), and prepend [Quoted: ...] to both content text and the first
content part. LRU-cached (200) per message_id.
- AbstractChannelAdapter gains getConfigLong helper for numeric config keys.
Frontend:
- types/index.ts feishu fields: default connection_mode is websocket; the
recommended option moves to the top; verification_token and encrypt_key
get showIf so they only render in webhook mode; new enable_quoted_context
switch (default on) exposes the quoted-message feature.
- ChannelEditModal builds a feishu-specific WEBHOOK_GUIDES path that picks
webhookStep vs websocketStep based on connection_mode, so users only see
steps for the mode they're using.
- i18n: split feishu.step3/step4 into webhookStep/websocketStep, rename
step5 to permissionStep. Channel type labels in zh-CN drop bilingual
prefix (e.g. 'Feishu / Lark (飞书)' -> '飞书').
Migrations:
- V52 was a no-op the first time it ran (matched compact JSON only) and
Flyway refused to re-run after the SQL was fixed. V52 is documented as a
no-op; V53 carries the actual UPDATE with REPLACE covering both compact
and pretty-printed JSON, and an idempotent WHERE for rows already on
websocket. h2 and mysql variants stay in lockstep.
Issue #24: tools selected in the agent binding UI had no effect at runtime.
mate_tool.name stores the Java class name (e.g. "BrowserUseTool") and was
written into mate_agent_tool.tool_name, but AgentToolSet.withAllowedToolsOnly
matched by the @Tool function name (e.g. "browser_use") — so every binding
was silently filtered out.
Fix: AgentToolSet builds an alias index per ToolCallback indexed by every
equivalent identifier — function name, Spring bean name, and Java class
simple name. withAllowedToolsOnly / withDeniedToolsFiltered / excluding
all accept any of these aliases, mirroring how Spring's BeanFactory accepts
bean names + aliases.
ToolRegistry.getEnabledToolSet now threads a bean→beanName resolver into
the new AgentToolSet.fromCallbacks(...) overload. Existing two-arg callers
keep working; tests pass without changes.
Zero data migration: stale mate_agent_tool rows that previously had no
effect now resolve correctly via the class-name alias.
- Reconcile approval status atomically: DB row, message metadata, in-memory store
- Approve and deny both flip the tool-call card + timeline segment to a terminal
state on the gate message — no more orange spinner stuck after a decision
- Frontend hydrate matches by pendingId and reverse-converges to expired so a
refresh after server-side timeout / consume clears the banner without restart
- Stop sweep, GC timeout, and JVM restart all close the loop with consistent
state
- Remove the dead REST /approve endpoint + matching frontend client export so
there is only one resolve path to maintain
Some self-hosted OpenAI-compatible serving frameworks return a 400 Bad Request
with a generic Pydantic "body=None / Field required" error when the outbound
request carries tool_choice="auto" but the server was launched without an
auto-tool-choice opt-in flag. The error message hides the real cause: the
request is rejected at validation time before the body is parsed, so the
upstream client sees only the generic body-missing error.
Per the OpenAI spec, omitting tool_choice when tools is non-empty is
functionally equivalent to "auto" — the server defaults to auto-pick.
Adding a stripAutoToolChoice patcher to the buildOpenAiApi chain:
- changes nothing on compliant servers (OpenAI / DashScope / DeepSeek / Kimi
default to auto when tools are present)
- unblocks strict OpenAI-compatible self-hosted endpoints
Explicit values other than "auto" ({"none", "required", or a function
descriptor}) are passed through unchanged.
Run on both chatCompletionEntity and chatCompletionStream paths so both
buffered and streaming calls benefit.
Three bugs surfaced when a non-admin workspace member opened the channel
admin page:
- vue-i18n "Invalid linked format" when '@' appeared in message strings
without the linked-format escape. Replaced literal '@' with vue-i18n v9
literal interpolation {'@'} in both zh-CN.ts and en-US.ts (6 strings:
QQ guide step3, accessControl requireMention/Tooltip).
- 403 from WorkspaceAccessInterceptor was being treated as 401 by the
axios interceptor and the chat SSE handler, clearing the token and
redirecting to /login. Split the two:
* 401 = authentication failure -> handleAuthFailure (logout)
* 403 = authorization failure -> keep session, surface to caller
Now a member who lacks workspace permission sees a toast instead of
being silently logged out.
- Two backend exception sites threw with the default code=500 for what
is semantically an auth/authz event, contradicting the codes returned
elsewhere for the same business event:
* AuthService.login() bad credentials 500 -> 401
* WorkspaceService.requirePermission() 500 -> 403
This aligns service-layer denials with SecurityConfig (401 for missing
JWT) and WorkspaceAccessInterceptor (403 for permission denied), so
the same business event always produces the same code.
Foundation for the ghost-approval root-cause fix.
Adds ResolveOutcome / MetadataDecision; rewrites ApprovalWorkflowService so
every resolve / consume / timeout / supersede transitions through one
two-phase contract: snapshot → DB UPDATE conditional on status=PENDING →
metadata reconciliation → afterCommit memory mutation. ChatController,
ChannelMessageRouter, and ApprovalController all switch to the workflow;
ApprovalService.resolve / resolveAndConsume / consumeApproved /
cancelStalePending / denyAllByConversation are physically removed so
DB-bypass is no longer reachable at compile time.
Specific fixes:
- recoverFromDb preserves DB pendingId + createdAt (was generating fresh
random ids, breaking every later DB sync)
- effectiveExpireAt = expireAt ?? createdAt + PENDING_TTL: legacy rows
with NULL expireAt no longer resurrect as live PENDING after restart
- markPendingApprovalsResolved flips pendingApproval.status + currentPhase
+ MessageEntity.status atomically (was only flipping the first field;
message.status uses existing completed/stopped, not approved/denied,
to stay within the frontend Message.status union)
- GC scheduler moves to ApprovalWorkflowService; timeouts and overflow
evictions now sync DB + metadata + memory through markTimeout
- DB UPDATE rows=0 returns alreadyResolved (concurrent-resolve safe);
exception propagates so @Transactional rolls back; memory stays untouched
- expireRecoveredRow gates metadata write on DB success (was writing
metadata even when DB update failed, producing the worst-case ghost)
- Mockito JDK 21 agent attach fixed via maven-dependency-plugin properties
+ surefire argLine (no more flaky self-attach across machines)
Tests: 34 new across 4 classes (recovery, resolve, GC, metadata sync).
Full suite: 788 / 788.
The two tools-sync scripts ran on every startup and used H2 MERGE INTO
... KEY(id), which overwrites every column on existing rows. That
silently reverted UI-toggled `enabled` and was the proximate cause of
a recent WriteFileTool/EditFileTool outage.
They were also a strict subset of the fresh-install seed (data-zh.sql /
data-en.sql register all 19 builtins; the sync scripts only 16) and out
of date. Per-tool Flyway migrations (V3, V31) are already the canonical
'register a new builtin' path, so the sync layer was duplicated and
error-prone.
Delete both files and the runToolSyncScript() loader. Tool descriptions
shown to the LLM come from @Tool annotations in code, not the DB row,
so removing per-startup metadata refresh has no functional impact.
Two follow-up improvements on top of renderDocxFromFile so the docx
pipeline can handle real long-form deliverables instead of just
prose-only memos.
Image embedding (P1).
MarkdownDocxRenderer now recognizes single-line  markdown
and embeds the referenced file via POI's XWPFRun.addPicture():
- PNG / JPG / GIF / BMP read straight from disk
- SVG rasterized via Apache Batik (PNGTranscoder, target width 1400px)
before embedding — OOXML stores raster images, so any vector source
needs conversion. Batik runs in-JVM, no rsvg-convert / cairo on host.
- Pictures are pinned to roughly the printable page width (≈ 5.77 in
for A4 minus default 1800-twip margins) and given a 4:3 height
fallback. Mixing images inline with other paragraph text is not
supported by design — the markdown subset assumes one image per
block paragraph. Inline images would require splitting paragraphs
across runs with explicit positioning, well beyond what this
renderer covers.
- Failure modes (missing file, unsupported format, Batik blowing up)
emit an italicised "[image: alt — reason]" placeholder so the rest
of the document still renders; the agent can read its own log to
see why the picture didn't make it.
- Adds two transitive deps via pom: batik-transcoder + batik-codec at
1.18, ~10 MB combined. Worth it given the alternative is shelling
out to system tooling.
Multi-file render (P2-lite).
New tool renderDocxFromFiles(List<String> filePaths, filename, pageSize)
reads several markdown files in order and renders one combined docx.
Lets the agent split a 30-page proposal into cover.md / ch1.md /
ch2.md / appendix.md and produce a single deliverable in one tool
call. Each path goes through WorkspacePathGuard.validatePath; any
empty or unreadable file aborts with a typed error so the agent
fixes its file list before retrying. Files are joined with a blank
line — no separator markup is injected, headings carry over cleanly.
I deliberately did NOT build the heavier mutable-docx state
("appendDocxChapter / finalizeDocx") flavor of P2: the multi-file
form covers the same workflow with no per-conversation state to
clean up, and the agent can iterate by rewriting the chapter file
and re-running the tool. Stateful append can come later if a
streaming use case actually shows up.
renderDocx and renderDocxFromFile @Tool descriptions updated to point
the agent at renderDocxFromFile for >5 KB markdown and to advertise
the new image-embedding capability.
renderDocx requires the markdown body to flow through the LLM as a
tool argument. For an 80 KB project proposal that's ≈ 20 K tokens of
streaming output spent just to repeat back content the model already
wrote to disk a turn earlier — multi-minute generation, real money.
renderDocxFromFile takes a file path instead. The agent uses
write_file / edit_file to assemble the markdown locally, then calls
this tool with just the path. JVM reads the file in one IO syscall
and feeds it to the existing MarkdownDocxRenderer. Token cost drops
from ≈ 20 K to ≈ 50 (the path string).
Behavior:
- Path resolution honors WorkspacePathGuard, same boundary as
read_file / write_file. No path traversal.
- UTF-8 read; rejects empty / missing / non-regular paths with
typed error messages so the agent can recover.
- Output cached in GeneratedFileCache and returned as a relative
/api/v1/files/generated/{id} link, with the same anti-host-
hallucination instruction renderDocx already carries.
- Same supported markdown subset (headings, bold, lists, tables).
Image references () still render as raw text — full
image embedding (P1) and SVG → PNG conversion (also P1) need
Apache Batik plus image-rendering plumbing in MarkdownDocxRenderer
and is tracked separately. Chapter-mode merge (P2) likewise needs
its own plumbing.
The @Tool description tells the agent to prefer this path when
markdown exceeds ~5 KB and shows the full write_file →
renderDocxFromFile workflow inline.
Three small but high-impact fixes that all surfaced together while
verifying the long-form generation flow.
1. ChatConsole onBeforeUnmount no longer kills the backend turn.
Previously, switching tabs / route navigation / any cause that
unmounted the chat view called stopChatGeneration(), which POSTs
/chat/{cid}/stop and aborts the in-flight LLM call. The user
reported a turn dying mid-generation just from switching pages.
Replaced with resetForNewConversation() — front-end SSE disconnect
only, no /stop. Backend keeps running; pollActivity / status probe
reconnects on return. Aligns with the existing comment in
selectConversation: "let A's backend agent run continue running."
2. Agent max_iterations raised 25 → 100 with a hard ceiling.
The previous 25-step ceiling caused LimitExceededNode to fire on
substantive multi-tool tasks (document generation + image conversion
+ retry loops). 100 matches QwenPaw's _MAX_MAX_ITERATIONS upper
bound. New plumbing:
- BaseAgent.MAX_ITERATIONS_HARD_CEILING = 100 public constant
- BaseAgent default field 25 → 100 (Java-side fallback)
- AgentGraphBuilder clamps any per-agent DB override to the
ceiling at runtime; if the row holds 200, runtime sees 100 and
a WARN is logged with the original value.
- V47 migration (h2 + mysql) idempotently bumps the three default
seeded agents (1000000001, 1000000002, 1000000003) only if they
still hold the old defaults (25 / 20). User-customized values
are not touched.
- data-en/zh/-mysql-en/-mysql-zh seed files updated to 100 for
fresh installs.
3. DocxRenderTool tells the LLM not to prepend a host to the URL.
DeepSeek and Claude have both been observed wrapping the
/api/v1/files/generated/{id} relative path returned by renderDocx
into an absolute URL with a hallucinated domain (e.g.
https://ai-tools-system.com/...), breaking the download link in
the rendered chat bubble. The tool's return string now appends an
explicit "must use the relative path verbatim, do not add any
https:// or http:// prefix" instruction, which Claude and
DeepSeek both honor.
A bundle of stability fixes that all surfaced together while running
the same long-form generation task across multiple turns. Each one
addresses a distinct way the previous behavior silently dropped
content the user had already seen on screen.
1. Mid-turn narrative persistence (StateGraphReActAgent +
SummarizingNode). Intermediate ReasoningNode rounds and
SummarizingNode broadcast their content_delta directly to the
SSE channel for live display, but the StreamAccumulator only
received the final answer. After refresh the assistant message
showed only tool_call cards with no body text.
StateGraphReActAgent now also forwards STREAMED_CONTENT (already
set per round) as a persistOnly StreamDelta whenever it changes,
so every narrative chunk lands in the accumulator's content
buffer and gets written to mate_message. SummarizingNode now
writes its summary into the same key so summarize narratives
persist too.
2. Follow-up message queue, not dispose (ChatController#interruptStream).
Sending a new message while a turn was running called
requestInterrupt, which dispose()d the active Reactor chain mid
LLM call. That cancelled the in-flight generation, lost partial
tokens, and left the user staring at a half-finished bubble.
The endpoint now uses enqueueMessage in all paths, matching
the "wait for current turn, then run" behavior. The old
requestInterrupt API is kept for any future force-replace UI
but no caller routes to it.
3. Queued user message ordering (ChatStreamTracker.QueuedInput +
ChatController.startQueuedMessage). interruptStream used to save
the queued user message immediately, before the in-flight
assistant message finalized in doOnError. listMessages orders
by create_time ASC, so the queued user message ended up above
the assistant reply it was supposed to follow. QueuedInput now
carries contentParts; persistence is delayed to startQueuedMessage,
which runs only after Asst-N is on disk.
4. JVM shutdown flush (ChatStreamTracker @PreDestroy +
emergencySaveAccumulator). A mvn spring-boot:run restart used to
wipe in-flight turns: SSE emitter timed out, ShutdownHook fired,
HikariPool closed before doOnError could save. ChatStreamTracker
now exposes an emergency-save callback per RunState; ChatController
registers one per stream that snapshots the accumulator and
writes status="interrupted_shutdown". @PreDestroy walks active
runs, invokes the callback, then disposes. Spring's reverse-order
bean teardown keeps ConversationService and Hikari alive long
enough for the save to complete.
5. Observation thresholds for summarize (GraphObservationProperties +
application.yml). The previous total-chars threshold of 12 KB
triggered summarize after one or two RFC reads, costing a 40 to
80 second compaction LLM call per loop. Tuned to: total 200 KB,
single 16 KB, large-result 32 KB, rounds safety net 25. Java
field defaults reverted to the conservative original values so
application.yml stays the source of truth.
6. Frontend thinking segmentation (useChat.ts thinking_delta +
phase). Multi-round ReAct turns merged every reasoning + summarize
round's thinking into one segment, accumulating to 9 KB+ in a
single bubble. thinking_delta now uses findLast(running) so a
tool_call_started or phase transition closes the previous segment
and the next delta opens a fresh one. phase event also closes
running thinking/content segments.
7. Other small things bundled: removed a debug metadata-keys log
that flooded the log file with one line per stream chunk; fixed
three stale tests that didn't compile after earlier constructor
changes (WikiLogServiceTest, WikiOverviewSpliceTest,
WikiProcessingServiceLazyTest); added rfc-066 documenting the
unified message queue + priority refactor as the next logical
step on top of these stabilizations.
Verified end-to-end with multiple full sessions: a four-minute
generation that produced the expected docx and a follow-up enqueue
that ran cleanly after the previous turn naturally completed,
without the old "Disposable unavailable" interrupt path.
Same bug as the prior queue-drop fix in doOnComplete, but in the
sister branch that fires when the agent's reactive stream errors
out (CancellationException from a user stop). The guard
cr.queuedInput() != null && !(isUserStop && !isInterruptFollowup)
mis-classified "user stopped, no interrupt-with-followup, but a
message is in the queue" as an explicit abort and silently dropped
the freshly-typed follow-up.
The frontend's enqueue path never sets interruptType — it just
calls requestStop + offers to messageQueue. Whoever puts a message
in the queue means it; just run it. Aligns with doOnComplete and
the four other queue-launch sites in this controller.
A series of cross-cutting stability fixes that surfaced together
during a long debugging session.
reasoning_content / Claude prefill self-replicating 400:
- ChatController persists typed errors (content starts with '[错误] ')
with status='error', so the failure text stops being re-sent as
multi-turn context — DeepSeek thinking 400 ('reasoning_content
must be passed back') and Claude 400 ('does not support assistant
message prefill') used to recursively re-create themselves every
retry by polluting history.
- BaseAgent.sanitizeForLlm filters status='error' / '[错误] ' prefix
assistant messages from history before LLM dispatch.
- BaseAgent.fetchHistoryMessages defensively drops trailing
AssistantMessages — Claude rejects assistant-tail prompts.
- NodeStreamingChatHelper.dropTrailingAssistant runs the same
defense at every doStreamCall pre-egress, so the in-turn
summarizing→reasoning transition (which leaves an assistant
scaffold at the tail) doesn't trip Claude either.
- AgentGraphBuilder.FallbackPolicy.DEEPSEEK switched (null,true,true)
→ (' ',false,true), aligning with KIMI/OPENAI's tolerant ' '
fallback. The previous 'force explicit 400' design was the
self-replicating loop's prime mover.
narration + tool args truncation:
- ReasoningNode.DEFAULT_MAX_OUTPUT_TOKENS 4096 → 16384. The 4k cap
was decapitating renderDocx tool_call args mid-stream when the
model emitted a long content field on top of thinking content;
the resulting 'invalid JSON' aborted execution silently.
- ReasoningNode appends a hermes-style TOOL_USE_ENFORCEMENT clause
to every system prompt: 'when you say you will perform an action,
call the tool now in the same response — narration is a protocol
violation'. Treats 'now I will generate the docx' (and never
actually calling renderDocx) as a forbidden pattern.
- ToolExecutionExecutor.normalizeToolExecutionError reframes the
JSON-truncated error as actionable instructions: 're-call the
same tool now with shorter content or split into multiple
sequential calls; do NOT describe the result as text'.
side fixes from the same evening:
- ChatController doOnComplete skips completionPublisher.publish
when isError=true, keeping memory extraction off the garbage path.
- ChatController doOnComplete queued-message guard simplified to
'cr.queuedInput() != null', matching the other 4 sites in the
controller. The previous 'isInterruptFollowup || !wasStopped'
guard silently dropped queued messages when the user did
Stop-then-Enqueue (wasStopped=true && interruptType=null), losing
the freshly-typed follow-up message.
- prompts/graph/summarize-system.txt now distinguishes 'single
task' (default; output one cohesive summary) from 'multiple
independent sub-tasks' (use the子任务 N format). Stops the
summarizer from inventing '子任务 1: PRO-027' decomposition for
unitary requests like 'write me a project proposal'.
- docker-compose.yml: pass SEARXNG_BASE_URL into mateclaw-server so the
app can reach the searxng sidecar container out of the box (default
http://searxng:8080).
- SystemSettingService: resolveSearxngBaseUrl() now falls back to the
SEARXNG_BASE_URL env var when no DB value is set, so Docker users no
longer need to configure it manually in the UI.
- V38 migration (h2 + mysql): expand mate_wiki_chunk.content from TEXT
(64KB) to MEDIUMTEXT (16MB) so large Chinese chunks (~30k chars
≈ 90KB UTF-8) no longer overflow.
- Add Node 22/pnpm stage to mateclaw-server/Dockerfile so docker compose
up -d --build produces a fully working image (the Vue SPA is now built
inside the image and copied into the JAR's classpath/static).
- Remove DASHSCOPE_API_KEY from .env.example and root docker-compose.yml;
LLM API keys are configured post-startup via the model management UI.
- mateclaw-server/settings.xml: Maven mirror routes requests through Aliyun
so dependency:go-offline no longer hangs in restricted networks
- Dockerfile: COPY settings.xml into /root/.m2/ before any mvn command
- docker-compose.yml: fix port mapping 18080->18088 (app listens on 18088)
- docker-compose.yml: remove legacy schema.sql/data.sql MySQL mounts;
Flyway manages schema creation from V1 baseline on startup
Five-commit bundle brings the Dream v2 P1 engine layer online, sitting
on top of the lifecycle mediator foundation already merged.
B.1-B.4 · Schema + records
- Flyway V26 (dream_report) + V27 (memory_recall review fields),
both h2 and mysql
- DreamReportEntity + DreamMode + DreamStatus enum + record types
- DreamReportMapper repository layer
B.5-B.8 · Consolidate refactor + focused dream
- MemoryEmergenceService refactored for plug-in dream modes
- MemoryRecallService extended with promoted/rejected review fields
- Focused dream endpoint + prompt template
- MemoryController exposes the review/trigger surface
B.9-B.10 · Monthly archive service
- MemoryArchiveService rolls cold promoted entries into archival rows
and reclaims daily_count storage
- DreamingScheduler runs archive job on its own schedule
B.12-B.14 · Tests
- MemoryArchiveServiceTest
- DreamFlagGuardTest
- DreamV2AcceptanceIT (end-to-end acceptance under feature flag)
Plus a verification script + HTTP e2e kit in the private test/ dir,
used for local staged rollout — not part of the open-source
distribution.
All features stay gated behind the mate.memory.dream.* flags from
Phase 1. Enable per-phase after staging validation.
beforeLlmCall / afterLlmCall / onSessionEnd only logged on failure,
making flag on/off indistinguishable in logs. Add debug lines on the
success path so lifecycle activation is observable.
Wire memory-facing events (turn-started, turn-completed, session-ended,
memory-written) through a single MemoryLifecycleMediator so
MemoryProvider implementations can hook into the agent conversational
flow without spreading side-effects across the runtime.
Ten atomic steps shipped under feat/dream-v2-p1-lifecycle:
- A.1 + A.2: MemoryLifecycleMediator class + TurnContext value object
- A.3: TurnStartedEvent / TurnCompletedEvent domain events
- A.4: MemoryLifecycleEventListener bean for Spring event plumbing
- A.5: MemoryProvider.onMemoryWrite default method (backward compatible)
- A.7: wire the mediator into AgentService at the right hook points
- A.8: LifecycleFlagGuardTest — feature flag must gate every hook
- A.9: MemoryLifecycleMediatorTest — unit coverage per hook
- A.10: LifecycleRecallCountIT — F4 regression across the stack
Feature flags (all default OFF; enable per phase after staging):
- mate.memory.lifecycle-mediator-enabled
- mate.memory.dream.focused-enabled
- mate.memory.dream.archive-enabled
This is Phase 1 foundation only — focused-dream and archive-dream
providers arrive in later phases.
Root cause: processRawMaterial() created a job record at queued stage
but never called jobService.transition() during processing. The job row
stayed at queued forever, so the stage bar never advanced.
Backend (WikiProcessingService):
- Transition job to ROUTING immediately after creation
- Transition to PHASE_A_RUNNING before chunk processing begins
- Transition to COMPLETED/PARTIAL/FAILED at the end based on finalStatus
- Transition to FAILED in the catch block on unhandled exceptions
Backend (WikiProcessingJobService.transition):
- Handle FAILED, PARTIAL terminal stages (set finishedAt + status)
- Handle non-terminal intermediate stages (set status to running)
Frontend (JobStageBar.vue):
- Add stageMapping for backend stages not shown as dots: phase_a_done →
phase_b_running, failed/partial/cancelled → completed position
- Guard stageIndex() against -1 (unknown stages default to all-pending)
- Terminal failure states show red failed dot instead of pulsing active
Two related changes that align buildFallbackChain with how users actually
think about failover.
1) Source = configured providers (was: only providers with fallback_priority > 0)
Earlier the chain was strictly "providers the user explicitly opted in via
fallback_priority > 0". A healthy in-pool provider with priority=0 was
silently excluded — surprising since the pool was supposed to be the source
of truth for "what is usable". After this change:
- Candidates = every configured provider
- Pool gating = same as before (in-pool members only at build time;
runtime walker re-checks)
- Order = agent prefs (PR-3) → fallback_priority asc (>0) →
priority==0 alphabetical
So fallback_priority is now purely an ordering hint, never an exclusion.
2) Per-provider model picker = default OR first-enabled (was: default only)
Previously a provider was skipped if no chat model on it had is_default=true.
That is admin friction with no benefit — every provider had to be visited in
Settings just to mark a default before it could appear in failover. New
pickFallbackModel():
- first try getDefaultModelByProvider — user explicit pick wins
- otherwise take the first enabled chat model on the provider
- skip only if neither exists
User-visible effect on the deployment that surfaced this:
- kimi-code primary fails (401 — real auth issue, separate from this bug)
- Pool short-circuits primary → walker fires
- Walker now sees dashscope (in-pool) AND ollama (in-pool) as candidates,
even though neither has fallback_priority set
- dashscope first enabled qwen model is picked → request succeeds via
dashscope without anyone touching Settings
45 failover-related tests still green (unit-level chain-build behavior is
backward-compatible; only the candidate set and model-selection lookups
changed, both broadening the chain rather than narrowing it).
Two real bugs the user restart surfaced — both turned healthy providers
into HARD-removed false positives.
Bug #1 — URL duplication
OpenAiCompatibleListModelsProbe always concatenated /v1/models, so
providers whose Base URL already includes the version segment got the
wrong URL:
LMStudio http://localhost:1234/v1 → /v1/v1/models → 404
ZhipuAI .../api/paas/v4 → /v4/v1/models → 404
Fix: detect a trailing /vN suffix and append /models instead. Six unit
tests in OpenAiCompatibleListModelsProbeTest lock the rule down.
Bug #2 — 404 false positives
Kimi for Coding API does not expose /v1/models even though chat works
fine, so the probe correctly received a 404 and incorrectly HARD-removed
the provider from the pool. Other vendors will hit the same — listing
is not a universal contract.
Fix: classify HTTP responses semantically.
401 / 403 → HARD remove (real auth failure)
404 / 405 / 410 → fail-open (endpoint missing, server may be alive)
other 4xx / 5xx → fail-open (probe inconclusive — let chat decide)
network errors → fail (unreachable)
This is the same philosophy as ChatGPTOAuthStatusProbe: when we cannot
cheaply confirm health, we do not proactively penalize the provider.
Same logic applied to Anthropic + DashScope probes for consistency.
Net effect on the user deployment after restart:
- kimi-code stays in pool (404 → fail-open) → primary path works again
- lmstudio + zhipu-cn also stay in pool (URL bug fixed)
- dashscope + ollama unchanged (real 200 OK)
Tests: 6 new for resolveModelsPath. The 2 unrelated WikiRawMaterialDedupTest
failures pre-date this commit and live in ba86bea.
Root cause: addFile()/addText() hash dedup only matched rows with
status=completed, so the same file uploaded while in partial/pending/
processing/failed status would create a duplicate row.
Fix:
- Remove .eq(processingStatus, "completed") from dedup queries — match
any non-deleted row with the same content hash in the KB
- On dedup hit: completed/pending/processing → return as-is;
partial/failed → trigger reprocess (partial enters resume branch)
- Clean up the newly uploaded temp file when dedup discards it
- Frontend: uploadRawFile/addRawText check for existing id in the list
before unshift to prevent visual duplicates
Test: WikiRawMaterialDedupTest — 10 cases covering all 5 statuses,
reprocess triggers for partial/failed, no-op for others, insert only
when no match.
PR-0 only installed the strategy seam; the actual ~600 LOC of provider-
specific construction stayed in AgentGraphBuilder as transitional public
helpers. PR-0b moves the DashScope + Anthropic halves into their builders
proper. (OpenAI larger refactor — 5 sub-helpers including Kimi/o-series
special cases — is left for a follow-up PR-0c.)
AgentDashScopeChatModelBuilder now owns:
- buildDashScopeApi (with provider/env/reflection key+url fallback chain)
- buildDashScopeOptions (model/temp/max-tokens/topP + built-in search)
- normalizeDashScopeBaseUrl (strip /compatible-mode/, return null for SDK default)
- readApiKeyFromDefaultChatModel + readBaseUrlFromDefaultChatModel +
readDashScopeApiFromDefaultChatModel (reflection-based final fallback)
- isBuiltinSearchEnabled (renamed from isDashScopeSearchEnabled, called
by AgentGraphBuilder.build via the now-injected dashScopeBuilder ref)
AgentAnthropicChatModelBuilder now owns:
- buildAnthropicApi (key validation, applyHttpTimeouts duplicated locally)
- buildAnthropicOptions (extended-thinking budget mapping low/medium/high/max
→ 4k/8k/16k/32k, temperature=1 enforcement, RFC-014 prompt cache options)
AgentGraphBuilder dropped:
- DashScope: ~120 LOC (api + options + 4 helpers + isDashScopeSearchEnabled)
- Anthropic: ~75 LOC (api + options)
- DashScopeChatModel + DashScopeConnectionProperties fields (unused after move)
- Deprecated single-fallback buildFallbackModel (no callers, superseded
by buildFallbackChain since RFC-009 PR-1)
- 5 imports for moved DashScope/Anthropic types
Net: -154 LOC in AgentGraphBuilder (1721 → 1567), +372 across the two new
builders. Strategy seam is now real for 3 of 4 protocols (ChatGPT was
already standalone, OpenAI is PR-0c). 220/220 tests still green — no
behavior change.
Two related issues from the Kimi-401 user report:
1. Backend (NodeStreamingChatHelper): a primary AUTH_ERROR (e.g. Kimi 401
with an invalid API key) returned immediately without trying the
fallback chain — a fallback provider with a different, valid key
never got a chance. Even with DashScope correctly configured as the
fallback, the user chat dead-ended on a 401.
The original assumption ("auth never self-heals so do not retry")
holds for the primary same-model retry loop but is wrong for the
fallback chain — different providers have different keys. Apply the
same break-into-fallback policy that BILLING and MODEL_NOT_FOUND
already use. recordPrimary(false) is preserved so the cooldown
counter still accumulates.
2. Frontend (chatError.ts + i18n): the error-text matching for
/认证|auth|unauthorized|401/i was so broad it matched the substring
"auth" inside URLs like https://api.kimi.com/.../auth, classifying
any model 401 as user "session expired" and rendering the misleading
"页面将自动跳转到登录页" copy. (The redirect itself only fires from
/api/v1/auth/* axios paths and SSE-connection 401s, not from this
payload-text path — but the copy alone is the worst kind of false
alarm.)
Add a new ChatErrorCategory provider_auth_error and split the
pattern matching: narrow auth_expired (HTTP 401 / 登录已过期 /
session expired / 凭证失效) is matched FIRST, then the broad
401-ish pattern routes to provider_auth_error. BACKEND_ERROR_TYPE_MAP
for AUTH_ERROR is also remapped, since structured backend payloads
currently always come from LLM providers — never from our own
/api/v1/auth path.
Tests
- NodeStreamingChatHelperFailoverTest (5 cases): primary 401 →
fallback succeeds; chain skips auth-failing fallback to next healthy
one; whole-chain failure surfaces last AUTH_ERROR (no silent drop);
BILLING regression unchanged; primary-success path does not touch
chain
- Browser preview verified: new i18n keys resolve in en-US, classifier
correctly routes "[错误] 401 from kimi.com" → provider_auth_error
while "[错误] HTTP 401 from /api/v1/auth/ping" stays auth_expired
- 186 tests pass (was 181 + 5 new); vue-tsc clean
Do-not-touch list: handleAuthFailure() in useStream/api/index.ts (real
session-expiry path) is unmodified — only the misclassification
upstream is fixed. auth_expired i18n copy is unchanged.
Track the primary model health, not just fallback entries
- NodeStreamingChatHelper accepts primaryProviderId via a new 5-arg
constructor; AgentGraphBuilder passes ModelConfigEntity.getProvider()
- Before the 5-retry primary loop, check
healthTracker.isInCooldown(primaryProviderId): if true, log + broadcast
"主模型暂时不可用（冷却中），直接尝试备选模型..." and short-circuit
straight to the fallback chain. Prevents a degraded primary from
burning 30+ seconds of backoff on every conversation turn.
- recordPrimary(success/failure) now fires on every primary verdict —
AUTH, BILLING, MODEL_NOT_FOUND, EMPTY_RESPONSE, generic UNKNOWN, and
the explicit success path. Three consecutive failures push the
primary provider into cooldown automatically.
- Legacy 1/2/3-arg constructors leave primaryProviderId null; tracking
silently disables for them so existing tests/wiring keep working.
Split BILLING and MODEL_NOT_FOUND out of CLIENT_ERROR / AUTH_ERROR
- BILLING (HTTP 402, "insufficient_quota", "credit balance is too low",
"billing_hard_limit_reached", "quota exceeded"): payment failure on
primary does not kill the call — a different provider may have credits.
Skips same-model retries and heads to fallback chain.
- MODEL_NOT_FOUND (HTTP 404, "Model not exist", "model_not_found",
DashScope "[InvalidParameter] url error"): unknown model id will not
start working on retry. Was previously misclassified as CLIENT_ERROR
and terminated the whole call; now routes to fallback so a different
provider can attempt with its default model.
- classifyError ordering matters: BILLING / MODEL_NOT_FOUND are matched
BEFORE the generic 400 / Bad Request branch, otherwise they would be
swallowed by CLIENT_ERROR.
Tests
- ErrorClassificationTest: 11 tests, covers multi-vendor error phrasing
for both new types + regression checks that 401 / 429 / 400 still
classify as before
- NodeStreamingChatHelperFallbackChainTest: +2 tests verifying
primaryProviderId persistence on the new constructor and null on
legacy ones
- 181 tests pass (was 168 + 13 new)
UI — Failover priority editor
- ProviderConfigRequest + ProviderInfoDTO carry fallbackPriority
- ModelProviderService.updateProviderConfig persists it (null = unchanged);
toProviderInfo exposes the current value to the UI (defaults to 0)
- ProviderConfigModal advanced panel exposes a number input with hint
- ProviderCard shows a "Fallback #N" badge for chain members so the
priority order is visible at a glance without opening the modal
- 5 new i18n keys (zh + en) — verified to resolve at runtime via i18n.global.t
Backend — Per-provider health tracker
- ProviderHealthTracker: ConcurrentHashMap-backed counters; N consecutive
failures (default 3) push the provider into a cooldown window (default
5 min) during which the chain walker skips it. Success resets both
counter and cooldown atomically. Lazy expiry on lookup so dead entries
do not accumulate.
- ProviderHealthProperties exposed under mateclaw.llm.failover.health.*
with sane production defaults
- New FallbackEntry record (providerId + ChatModel) replaces raw
List<ChatModel> in the chain so the walker can correlate cooldown
state to entries; AgentGraphBuilder.buildFallbackChain returns the
new type
- NodeStreamingChatHelper takes the tracker through a new 4-arg
constructor and consults it before each fallback call; records
success/failure on each chain attempt. Legacy 2/3-arg constructors
preserved as @Deprecated wrappers (synthetic providerId means no
health tracking on the legacy path — that path is opt-out anyway)
Tests
- ProviderHealthTrackerTest (9 tests): below/at threshold, success
reset, cooldown expiry (via reflection on the min-clamp setter),
disabled-tracker no-op, null-providerId safety, per-provider
isolation, snapshot output
- NodeStreamingChatHelperFallbackChainTest updated to FallbackEntry
field type — verifies providerId + ChatModel survive the chain
- 168 tests pass (was 159 + 9 new)
Verification
- mvn test green; vue-tsc clean; live UI confirms i18n resolution
Replaces the hardcoded single-DashScope fallback with a DB-driven
ordered chain. Same-provider primary deployments (e.g., DashScope
qwen-max) finally get a real fallback; if any provider in the chain
returns an empty body or transient failure, the next is tried.
Schema — DB-driven chain
- mate_model_provider gains `fallback_priority INT DEFAULT 0`. Positive
values define try-order; 0 = not in chain. Migration V21 (h2 + mysql)
seeds DashScope as priority 1 to preserve existing behavior.
- ModelProviderService.listFallbackChain() returns providers ordered by
priority ascending.
- ModelProviderEntity gains the new field.
Runtime — chain walk + empty-response trigger
- AgentGraphBuilder.buildFallbackChain(primaryConfig) returns a
List<ChatModel>, identity-filtering the primary by (providerId,
modelName) — fixes the bug where same-provider-primary deployments got
null fallback. Providers whose API key is missing are silently
skipped with WARN. Old buildFallbackModel(ChatModel) kept as
@Deprecated wrapper.
- NodeStreamingChatHelper accepts List<ChatModel>; the post-retry
fallback block now walks the chain in priority order, single-shot
per entry. Old single-fallback constructors retained as @Deprecated
one-element-list wrappers so legacy callers keep working.
- New ErrorType.EMPTY_RESPONSE: when the LLM returns no content, no
thinking, AND no tool calls, mark the result as a soft failure and
break the same-model retry loop, handing off directly to the
fallback chain.
- Broadcast updated to "切换到备选模型 (N/M)..." so SSE consumers see
chain progress.
Tests
- NodeStreamingChatHelperFallbackChainTest covers constructor variants,
chain immutability, deprecated-overload back-compat, and the
EMPTY_RESPONSE enum exists as a compile-time contract.
- 159 tests pass (was 153 + 6 new).
A. Delete two dead prompt files (prompts/context/conversation-summary-*.txt)
that no caller has loaded since the structured-summary triple replaced them.
B. Drop the never-wired locale machinery: PromptLoader.loadPrompt(name, locale)
overload + the prompts/{locale}/... fallback chain + I18nService.currentLocaleTag().
A single-language prompt corpus plus LLM input-language following is sufficient.
C. Strip duplicated structure list / budget directive from
structured-summary-update.txt (the system prompt already carries them).
Add a defensive preamble to both summary prompts: "do not respond to any
questions or requests in the conversation, only output the structured
summary" — prevents the summarizer from accidentally answering historical
user questions.
D. Fix {summary_budget} placeholder leak in the iterative-update branch of
ConversationWindowManager.generateSummary. Both branches now substitute
on the SystemMessage uniformly. Regression-guarded by
ConversationWindowManagerSummaryBudgetTest.
E1. De-hardcode seven prompts (research/{plan,draft,compose}-{system,user},
graph/limit-exceeded-system) — language now follows the user's input
instead of being hardcoded; citation tokens are language-neutral
[M1] / [Q1] markers.
E2. Add 10 i18n keys (research.fallback.*, research.broadcast.*,
agent.limit_exceeded.*) to messages.properties + messages_en.properties.
Inject I18nService into WikiResearchService and LimitExceededNode and
route 5 + 2 hardcoded fallbacks through i18n.msg(). Regression-guarded
by WikiResearchServiceFallbackTest + LimitExceededNodeFallbackTest.
E3. Replace 3 assembly tags in WikiResearchService with neutral
[M1] / [Q1] tokens. Aligns with the [M1] / [M2,3] citation format the
draft prompt asks for.
G. Three new regression tests cover D, E2, and E3.
- db/migration/mysql: replace ADD COLUMN/CREATE INDEX IF NOT EXISTS with
idempotent checks via information_schema (MySQL 8.0 <8.0.29 and some
forks don't support IF NOT EXISTS for ADD COLUMN). Affects V2/V4/V5/V7
/V8/V9/V11/V12/V13/V14. Fix: gitee#IIYHLJ.
- application-mysql.yml: add createDatabaseIfNotExist=true so MySQL
Connector/J auto-creates the schema on first connection (requires
CREATE privilege — documented fallback for restricted accounts).
- llm/OllamaAutoDiscoveryRunner: rewrite seed tag when fuzzy-matching,
prefer exact tag for default; skip models without tool support when
auto-activating a default (prevents the phantom ':latest' trap when
users pulled a specific size).
- agent/graph/NodeStreamingChatHelper: detect 'does not support tools'
and 'model not found' errors from Ollama and emit actionable Chinese
prompts guiding users to qwen3 / qwen2.5:7b+ / llama3.1:8b+ etc.
- ui/MessageBubble + types/chatError: surface the backend's actionable
rawMessage in the failed-message card instead of a generic '未知错误';
strip redundant prefixes (Bad request: / [错误] / LLM 调用失败:) since
the title already conveys the category.
- ChannelMessageRouter: include assistantMessageId in message_complete
and done broadcasts so ChatConsole observers can reconcile the
streaming placeholder to the persisted DB row by id instead of
falling back to the FIFO 'claim' heuristic (which occasionally
dropped the assistant bubble on external channel conversations).
Capture the id from ConversationService.saveMessage in both the
sync agentService.chat path and the streaming processWithStreaming
path; switch to HashMap since Map.of rejects null values when save
is skipped (e.g. under approval).
- ChatConsole: add a 'running' indicator on the sidebar so users can
tell which conversations have an in-flight agent run. Pulsing amber
dot on the channel icon (both expanded and collapsed modes) plus a
'生成中…' / 'Generating…' pill in expanded mode.
- ChatConsole: don't cancel the previous conversation's streaming run
when switching conversations — let it keep running in the background
and reconcile when the user comes back.
- ConversationWindowManager: cap reserve token at 50% of effective max
to prevent negative historyBudget on small-context models (8K/16K)
- common.security.SecretEquals: new constant-time comparison utility
(MessageDigest.isEqual wrapper) for secrets/tokens/signatures
- WeixinChannelAdapter: migrate context_token comparison to SecretEquals
- FeishuChannelAdapter: fail-fast on empty encrypt_key when connection_mode=webhook
- TelegramChannelAdapter: sanitize attachment captions — strip control bytes
(\p{Cc} except \t\r\n) + format chars (\p{Cf}) + 4096 char cap
- AgentGraphBuilder: fallback Anthropic max_tokens to 4096 on null/0/negative
Tests: SecretEqualsTest (5) + TelegramCaptionSanitizeTest (5) — all green.
