lpf
|
5e421bb730
|
Tighten subagent prompt file workflow
|
2026-03-06 14:12:01 +08:00 |
|
lpf
|
2bc71870db
|
Refine agent config schema and prompt file loading
|
2026-03-06 13:56:38 +08:00 |
|
lpf
|
6902f65c54
|
Add multi-agent config and registry runtime flow
|
2026-03-06 12:47:05 +08:00 |
|
lpf
|
3894f0e27a
|
fix pipeline tool context origin injection
|
2026-03-05 22:41:21 +08:00 |
|
lpf
|
b9e1efa3ef
|
fix subagent completion notify path and pipeline origin routing
|
2026-03-05 22:17:44 +08:00 |
|
lpf
|
e8e1cdae32
|
feat: ship subagent runtime and remove autonomy/task legacy
|
2026-03-05 21:54:54 +08:00 |
|
lpf
|
29d6480058
|
feat: add subagent profiles, memory namespaces, and webui management
|
2026-03-05 21:05:29 +08:00 |
|
lpf
|
94e8f2c9d9
|
fix
|
2026-03-05 12:55:30 +08:00 |
|
lpf
|
2fbb98bccd
|
feat: harden concurrency scheduling and task watchdog
|
2026-03-05 11:32:06 +08:00 |
|
lpf
|
09c06786a4
|
fix: harden telegram streaming finalize and resolve session lock-copy vet issue
|
2026-03-04 14:16:48 +08:00 |
|
lpf
|
b007a403a1
|
feat: improve telegram streaming stability and task context with memory/ekg
|
2026-03-04 14:05:47 +08:00 |
|
lpf
|
a7dcafae60
|
refactor: default-enable session planning and scheduling without config flags
|
2026-03-04 12:49:43 +08:00 |
|
lpf
|
154ab3f7f9
|
feat: add session auto-planning and resource-based concurrency scheduling
|
2026-03-04 12:44:31 +08:00 |
|
lpf
|
ba8cfbe131
|
fix(telegram-stream): chunked streaming with retry/rate-limit and suppress duplicate final send
|
2026-03-04 12:00:25 +08:00 |
|
lpf
|
d639a5113f
|
fix config
|
2026-03-04 11:04:25 +08:00 |
|
lpf
|
6c78ef3b1c
|
refactor(responses): remove chat_completions and wire responses features end-to-end
|
2026-03-04 10:25:50 +08:00 |
|
lpf
|
bd93c12edc
|
fix logs
|
2026-03-03 10:36:53 +08:00 |
|
野生派Coder～
|
e85bf4c5b8
|
Refine cron outbound fix with smaller implementation
|
2026-03-02 22:49:20 +08:00 |
|
野生派Coder～
|
67dd45a4dd
|
fix(cron): ensure reminders keep delivery channel target
|
2026-03-02 22:25:18 +08:00 |
|
|
|
d2f680debb
|
task lifecycle + heartbeat resilience: preserve archived done tasks and reset heartbeat session on pairing-loop errors
|
2026-03-02 10:31:31 +00:00 |
|
|
|
ed47e2dfe0
|
agent system-message: rebuild context and retry after purging orphan tool outputs on provider pairing 400
|
2026-03-01 15:06:46 +00:00 |
|
|
|
6608456fbf
|
task-audit cleanup: filter heartbeat by default and merge recurring autonomy runs via stable task ids; add channel-level inbound dedupe
|
2026-03-01 05:33:47 +00:00 |
|
|
|
78e3179135
|
ekg p0: configurable threshold, auto-repair task on repeated errsig, and provider/model audit visibility
|
2026-03-01 04:43:21 +00:00 |
|
|
|
ed9e4203d3
|
ekg m2+: add error-signature-aware provider ranking for fallback selection
|
2026-03-01 04:32:34 +00:00 |
|
|
|
1d5454066a
|
ekg m2: rank fallback providers by recent success/error history and record provider outcomes
|
2026-03-01 04:27:24 +00:00 |
|
|
|
0e8d267d78
|
ekg m1: add execution knowledge graph design + error-signature advice to suppress autonomy retry loops
|
2026-03-01 04:20:24 +00:00 |
|
|
|
c5033e8484
|
provider fallback: switch to proxy_fallbacks on API failure in message/system loops
|
2026-02-28 15:06:55 +00:00 |
|
|
|
058b215704
|
autonomy M1 phase2: expose idle-run source and block reasons in task queue; add whitelist policy plumbing
|
2026-02-28 06:01:29 +00:00 |
|
|
|
46a0403b0f
|
media abstraction: add structured MediaItem model and expose media-source audit in task queue detail
|
2026-02-28 02:49:13 +00:00 |
|
|
|
f274acf818
|
task-queue: unified task audit events, queue API, and bilingual split-view webui
|
2026-02-28 02:41:05 +00:00 |
|
|
|
e289b1fc24
|
agent audit phase3: disable progress pushes, keep final-result only and expose task audit API for webui
|
2026-02-28 02:25:26 +00:00 |
|
|
|
ca4f78c627
|
agent async-loop phase2: add task-id progress notices and completion correlation
|
2026-02-28 02:16:04 +00:00 |
|
|
|
4dcfa92717
|
agent async-loop: add long-run progress notices while keeping final result delivery
|
2026-02-28 02:08:31 +00:00 |
|
|
|
198bc8579f
|
agent long-run phase3: remove fixed tool-iteration cap by default, keep configurable ceiling
|
2026-02-28 02:02:34 +00:00 |
|
|
|
bd7ad72550
|
agent long-run phase2: configurable hard cap for tool iterations and safer final result fallback
|
2026-02-28 01:11:36 +00:00 |
|
|
|
9feb4d0c53
|
normalize-policy phase-4: remove extra trim/lower from language/trigger/context handling
|
2026-02-27 17:06:53 +00:00 |
|
|
|
4a9baf91ca
|
agent long-run: extend tool-chain iterations and force final synthesis to avoid premature fallback
|
2026-02-27 16:56:19 +00:00 |
|
|
|
a75be4b618
|
normalize-policy: preserve raw casing/spacing in action-content sensitive paths (phase-3)
|
2026-02-27 16:48:29 +00:00 |
|
|
|
4c2a08396e
|
agent: avoid no-response fallback after long tool runs by summarizing tool outputs
|
2026-02-27 09:41:24 +00:00 |
|
|
|
7bb2ad86ec
|
memory: append daily work summary entries after each completed response
|
2026-02-27 06:01:50 +00:00 |
|
|
|
d51fdb634c
|
p0 rollback: disable telegram token streaming by default
|
2026-02-26 15:52:49 +00:00 |
|
|
|
d38d4d2d26
|
agent: add non-empty fallback when model returns empty final content
|
2026-02-26 14:15:44 +00:00 |
|
|
|
36ea7486d1
|
p0: purge orphan tool outputs on pairing error to prevent repeated 400 loops
|
2026-02-26 13:45:41 +00:00 |
|
|
|
4f5fb0e6cf
|
telegram streaming: finalize placeholder when response suppressed
|
2026-02-26 12:55:53 +00:00 |
|
|
|
818408962d
|
streaming: implement provider→agent→telegram token streaming pipeline
|
2026-02-26 12:51:50 +00:00 |
|
lpf
|
06c7f2b53f
|
fix
|
2026-02-26 15:45:51 +08:00 |
|
lpf
|
1ae8d384bb
|
fix loop
|
2026-02-26 15:35:26 +08:00 |
|
lpf
|
5f31ae5c99
|
fix context
|
2026-02-26 14:07:32 +08:00 |
|
|
|
f8c662a468
|
fix: enforce strict tool-call pairing across chat paths
|
2026-02-26 05:47:35 +00:00 |
|
|
|
2d0ca1d4ce
|
session alignment: force single main session and openclaw-style agents/main/sessions path
|
2026-02-26 02:20:51 +00:00 |
|
