Commit Graph

3291 Commits

Author SHA1 Message Date
Peter Steinberger
17fd46ab66 test: fix websocket tool shape coverage 2026-03-12 02:16:56 +00:00
Peter Steinberger
980619b9be fix: harden openai websocket replay 2026-03-12 02:13:06 +00:00
Brian Yu
cced1e0f76 preserve openai phase param 2026-03-11 23:15:52 +00:00
zhoulf1006
453c8d7c1b fix(hooks): add missing trigger and channelId to agent_end, llm_input, and llm_output hook contexts (#42362)
Merged via squash.

Prepared head SHA: e6d7b7e31aa6ae12813b4609adb6e569a4084d08
Co-authored-by: zhoulf1006 <35586967+zhoulf1006@users.noreply.github.com>
Co-authored-by: hydro13 <6640526+hydro13@users.noreply.github.com>
Reviewed-by: @hydro13
2026-03-11 23:40:13 +01:00
Gustavo Madeira Santana
d79ca52960 Memory: add multimodal image and audio indexing (#43460)
Merged via squash.

Prepared head SHA: a994c07190a2062322f459c928b6cd74f9803d88
Co-authored-by: gumadeiras <5599352+gumadeiras@users.noreply.github.com>
Co-authored-by: gumadeiras <5599352+gumadeiras@users.noreply.github.com>
Reviewed-by: @gumadeiras
2026-03-11 22:28:34 +00:00
Peter Steinberger
c8dd06cba2 fix(ws): preserve payload overrides 2026-03-11 20:11:51 +00:00
Peter Steinberger
bdd9ed238a test: align pi-ai oauth mocks 2026-03-11 20:11:51 +00:00
Peter Steinberger
620bae4ec7 fix(ollama): share model context discovery 2026-03-11 20:11:51 +00:00
Peter Steinberger
9329a0ab24 test(agents): cover openai responses phase replay 2026-03-11 20:10:55 +00:00
Tak Hoffman
4133edb395 fix: restore web tools to coding profile (#43436)
* fix: restore web tools to coding profile

* fix: tighten tool catalog regression assertion
2026-03-11 15:07:17 -05:00
Squabble9
128e5bc317 fix: recognize Venice 402 billing errors for model fallback (#43205)
Merged via squash.

Prepared head SHA: 1f6b10b9d934235e71f279f888292139c4a85aa6
Co-authored-by: Squabble9 <194720422+Squabble9@users.noreply.github.com>
Co-authored-by: altaywtf <9790196+altaywtf@users.noreply.github.com>
Reviewed-by: @altaywtf
2026-03-11 22:15:32 +03:00
ingyukoh
2a18cbb110 fix(agents): prevent false billing error replacing valid response text (#40616)
Merged via squash.

Prepared head SHA: 05179362b439ff3b0330df3eb6a28aa05396ea9d
Co-authored-by: ingyukoh <6015960+ingyukoh@users.noreply.github.com>
Co-authored-by: altaywtf <9790196+altaywtf@users.noreply.github.com>
Reviewed-by: @altaywtf
2026-03-11 22:00:11 +03:00
VibhorGautam
4473242b4f fix: use unknown instead of rate_limit as default cooldown reason (#42911)
Merged via squash.

Prepared head SHA: bebf6704d7b02b9a32935c0006eac2d76694fec0
Co-authored-by: VibhorGautam <55019395+VibhorGautam@users.noreply.github.com>
Co-authored-by: altaywtf <9790196+altaywtf@users.noreply.github.com>
Reviewed-by: @altaywtf
2026-03-11 21:34:14 +03:00
Bill Chirico
60aed95346 feat(memory): add gemini-embedding-2-preview support (#42501)
Merged via squash.

Prepared head SHA: c57b1f8ba2ca65f4946afe94a9137ee8c05c8c64
Co-authored-by: BillChirico <13951316+BillChirico@users.noreply.github.com>
Co-authored-by: gumadeiras <5599352+gumadeiras@users.noreply.github.com>
Reviewed-by: @gumadeiras
2026-03-11 14:28:53 -04:00
ademczuk
58634c9c65 fix(agents): check billing errors before context overflow heuristics (#40409)
Merged via squash.

Prepared head SHA: c88f89c462d87957a4c6c51a23ab997fd307059d
Co-authored-by: ademczuk <5212682+ademczuk@users.noreply.github.com>
Co-authored-by: altaywtf <9790196+altaywtf@users.noreply.github.com>
Reviewed-by: @altaywtf
2026-03-11 21:08:55 +03:00
Tak Hoffman
87876a3e36 Fix env proxy bootstrap for model traffic (#43248)
* Fix env proxy bootstrap for model traffic

* Address proxy dispatcher review followups

* Fix proxy env precedence for empty lowercase vars
2026-03-11 10:21:35 -05:00
Bruce MacDonald
d6108a6f72 Onboard: add Ollama auth flow and improve model defaults
Add Ollama as a auth provider in onboarding with Cloud + Local mode
selection, browser-based sign-in via /api/me, smart model suggestions
per mode, and graceful fallback when the default model is unavailable.

- Extract shared ollama-models.ts
- Auto-pull missing models during onboarding
- Non-interactive mode support for CI/automation

Closes #8239
Closes #3494

Co-Authored-By: Jeffrey Morgan <jmorganca@gmail.com>
2026-03-11 14:52:55 +00:00
Robin Waslander
62d5df28dc fix(agents): add nodes to owner-only tool policy fallbacks
The nodes tool was missing from OWNER_ONLY_TOOL_NAME_FALLBACKS in
tool-policy.ts. applyOwnerOnlyToolPolicy() correctly removed gateway
and cron for non-owners but kept nodes, which internally issues
privileged gateway calls: node.pair.approve (operator.pairing) and
node.invoke (operator.write).

A non-owner sender could approve pending node pairings and invoke
arbitrary node commands, extending to system.run on paired nodes.

Add nodes to the fallback owner-only set. Non-owners no longer receive
the nodes tool after policy application; owners retain it.

Fixes GHSA-r26r-9hxr-r792
2026-03-11 14:17:03 +01:00
Andyliu
10e6e27451 fix(models): guard optional model input capabilities (#42096)
Merged via squash.

Prepared head SHA: d398fa0222b7045b549fd3592d469c079ca3efb6
Co-authored-by: andyliu <2377291+andyliu@users.noreply.github.com>
Co-authored-by: hydro13 <6640526+hydro13@users.noreply.github.com>
Reviewed-by: @hydro13
2026-03-11 13:43:59 +01:00
Frank Yang
d68d4362ee fix(context-pruning): cover image-only tool-result pruning 2026-03-11 18:07:37 +08:00
MoerAI
a78674f115 fix(context-pruning): prune image-containing tool results instead of skipping them (#41789) 2026-03-11 18:07:37 +08:00
ademczuk
dc4441322f fix(agents): include azure-openai in Responses API store override (#42934)
Merged via squash.

Prepared head SHA: d3285fef41001bb25a8d1cb47a37ee9a132ffb9e
Co-authored-by: ademczuk <5212682+ademczuk@users.noreply.github.com>
Co-authored-by: frankekn <4488090+frankekn@users.noreply.github.com>
Reviewed-by: @frankekn
2026-03-11 16:16:10 +08:00
Luke
7761e7626f Providers: add Opencode Go support (#42313)
* feat(providers): add opencode-go provider support and onboarding

* Onboard: unify OpenCode auth handling openclaw#42313 thanks @ImLukeF

* Docs: merge OpenCode Zen and Go docs openclaw#42313 thanks @ImLukeF

* Update CHANGELOG.md

---------

Co-authored-by: Ubuntu <ubuntu@vps-90352893.vps.ovh.ca>
Co-authored-by: Vincent Koc <vincentkoc@ieee.org>
2026-03-11 01:31:06 -04:00
Vincent Koc
bd33a340fb fix(sandbox): sanitize Docker env before marking OPENCLAW_CLI (#42256)
* Sandbox: sanitize Docker env before exec marker injection

* Sandbox: add regression test for Docker exec marker env

* Sandbox: disable Windows shell fallback for Docker

* Sandbox: cover Windows Docker wrapper rejection

* Sandbox: test strict env sanitization through Docker args
2026-03-11 00:59:36 -04:00
Peter Steinberger
a52104c235 test: restore fs bridge helper export 2026-03-11 02:38:00 +00:00
Peter Steinberger
a0d5462571 fix(security): pin staged writes and fs mutations 2026-03-11 02:38:00 +00:00
Peter Steinberger
72b0e00eab refactor: unify sandbox fs bridge mutations 2026-03-11 02:10:23 +00:00
Peter Steinberger
aad014c7c1 fix: harden subagent control boundaries 2026-03-11 01:44:38 +00:00
Peter Steinberger
68c674d37c refactor(security): simplify system.run approval model 2026-03-11 01:43:06 +00:00
Peter Steinberger
11924a7026 fix(sandbox): pin fs-bridge staged writes 2026-03-11 01:15:47 +00:00
Peter Steinberger
ecdbd8aa52 fix(security): restrict leaf subagent control scope 2026-03-11 01:12:22 +00:00
Peter Steinberger
fa0329c340 test: cover cron nested lane selection 2026-03-11 00:02:00 +00:00
Josh Avant
36d2ae2a22 SecretRef: harden custom/provider secret persistence and reuse (#42554)
* Models: gate custom provider keys by usable secret semantics

* Config: project runtime writes onto source snapshot

* Models: prevent stale apiKey preservation for marker-managed providers

* Runner: strip SecretRef marker headers from resolved models

* Secrets: scan active agent models.json path in audit

* Config: guard runtime-source projection for unrelated configs

* Extensions: fix onboarding type errors in CI

* Tests: align setup helper account-enabled expectation

* Secrets audit: harden models.json file reads

* fix: harden SecretRef custom/provider secret persistence (#42554) (thanks @joshavant)
2026-03-10 23:55:10 +00:00
Peter Steinberger
0bac47de51 refactor: split tar.bz2 extraction helpers 2026-03-10 23:53:32 +00:00
Peter Steinberger
9c64508822 refactor: rename tar archive preflight checker 2026-03-10 23:52:51 +00:00
Peter Steinberger
658cf4bd94 fix: harden archive extraction destinations 2026-03-10 23:49:35 +00:00
David Guttman
9f5dee32f6 fix(acp): implicit streamToParent for mode=run without thread (#42404)
* fix(acp): implicit streamToParent for mode=run without thread

When spawning ACP sessions with mode=run and no thread binding,
automatically route output to parent session instead of Discord.
This enables agent-to-agent supervision patterns where the spawning
agent wants results returned programmatically, not posted as chat.

The change makes sessions_spawn with runtime=acp and thread=false
behave like direct acpx invocation - output goes to the spawning
session, not to Discord.

Fixes the issue where mode=run without thread still posted to Discord
because hasDeliveryTarget was true when called from a Discord context.

* fix: use resolved spawnMode instead of params.mode

Move implicit streamToParent check to after resolveSpawnMode so that
both explicit mode="run" and omitted mode (which defaults to "run"
when thread is false) correctly trigger parent routing.

This fixes the issue where callers that rely on default mode selection
would not get the intended parent streaming behavior.

* fix: tighten implicit ACP parent relay gating (#42404) (thanks @davidguttman)

---------

Co-authored-by: Onur Solmaz <2453968+osolmaz@users.noreply.github.com>
2026-03-10 21:42:15 +01:00
Peter Steinberger
283570de4d fix: normalize stale openai completions transport 2026-03-10 20:23:03 +00:00
Peter Steinberger
6d4241cbd9 fix: wire modelstudio env discovery (#40634) (thanks @pomelo-nwu) 2026-03-10 19:58:43 +00:00
Josh Avant
0687e04760 fix: thread runtime config through Discord/Telegram sends (#42352) (thanks @joshavant) (#42352) 2026-03-10 13:30:57 -05:00
Yufeng He
c2d9386796 fix: log auth profile resolution failures instead of swallowing silently (#41271)
Merged via squash.

Prepared head SHA: 049d1e119a4df88ae00870353a9e7134261fe9dd
Co-authored-by: he-yufeng <40085740+he-yufeng@users.noreply.github.com>
Co-authored-by: altaywtf <9790196+altaywtf@users.noreply.github.com>
Reviewed-by: @altaywtf
2026-03-10 20:38:49 +03:00
JiangNan
e9e8b81939 fix(failover): classify Gemini MALFORMED_RESPONSE as retryable timeout (#42292)
Merged via squash.

Prepared head SHA: 68f106ff49fc7a28a806601bc8ca1e5e77c6e8c6
Co-authored-by: jnMetaCode <12096460+jnMetaCode@users.noreply.github.com>
Co-authored-by: altaywtf <9790196+altaywtf@users.noreply.github.com>
Reviewed-by: @altaywtf
2026-03-10 20:34:32 +03:00
jiarung
bc9b35d6ce fix(logging): include model and provider in overload/error log (#41236)
Merged via squash.

Prepared head SHA: bb16fecbf7173dbd8f49adacb6147635bad00c56
Co-authored-by: jiarung <16461359+jiarung@users.noreply.github.com>
Co-authored-by: altaywtf <9790196+altaywtf@users.noreply.github.com>
Reviewed-by: @altaywtf
2026-03-10 20:32:14 +03:00
CryUshio
8bf64f219a fix: recognize Poe 402 'used up your points' as billing for fallback (#42278)
Merged via squash.

Prepared head SHA: f3cdfa76dd9afcb023504eef723036e826e6ebc5
Co-authored-by: CryUshio <30655354+CryUshio@users.noreply.github.com>
Co-authored-by: altaywtf <9790196+altaywtf@users.noreply.github.com>
Reviewed-by: @altaywtf
2026-03-10 20:17:36 +03:00
sline
bfeea5d23f fix(agents): prevent /v1beta duplication in Gemini PDF URL (#34369)
Strip trailing /v1beta from baseUrl before appending the version
segment, so callers that already include /v1beta in their base URL
(e.g. subagent-registry) no longer produce /v1beta/v1beta/models/…
which results in a 404 from the Gemini API.

Closes #34312

Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com>
2026-03-10 12:52:49 -04:00
Ayaan Zaidi
ac88a39acc fix: align pi-ai 0.57.1 oauth imports and payload hooks 2026-03-10 20:29:03 +05:30
joshavant
59bc3c6630 Agents: align onPayload callback and OAuth imports 2026-03-10 08:50:30 -05:00
George Zhang
309162f9a2 fix: strip leaked model control tokens from user-facing text (#42173)
Models like GLM-5 and DeepSeek sometimes emit internal delimiter tokens in their responses. Uses generic pattern in the text extraction pipeline, following the same architecture as stripMinimaxToolCallXml.

Closes #40020
Supersedes #40573

Co-authored-by: imwyvern <100903837+imwyvern@users.noreply.github.com>
2026-03-10 06:27:59 -07:00
Charles Dusek
048e25c2b2 fix(agents): avoid duplicate same-provider cooldown probes in fallback runs (#41711)
Merged via squash.

Prepared head SHA: 8be8967bcb4be81f6abc5ff078644ec4efcfe7a0
Co-authored-by: cgdusek <38732970+cgdusek@users.noreply.github.com>
Co-authored-by: altaywtf <9790196+altaywtf@users.noreply.github.com>
Reviewed-by: @altaywtf
2026-03-10 15:26:47 +03:00
Pejman Pour-Moezzi
aca216bfcf feat(acp): add resumeSessionId to sessions_spawn for ACP session resume (#41847)
* feat(acp): add resumeSessionId to sessions_spawn for ACP session resume

Thread resumeSessionId through the ACP session spawn pipeline so agents
can resume existing sessions (e.g. a prior Codex conversation) instead
of starting fresh.

Flow: sessions_spawn tool → spawnAcpDirect → initializeSession →
ensureSession → acpx --resume-session flag → agent session/load

- Add resumeSessionId param to sessions-spawn-tool schema with
  description so agents can discover and use it
- Thread through SpawnAcpParams → AcpInitializeSessionInput →
  AcpRuntimeEnsureInput → acpx extension runtime
- Pass as --resume-session flag to acpx CLI
- Error hard (exit 4) on non-existent session, no silent fallback
- All new fields optional for backward compatibility

Depends on acpx >= 0.1.16 (openclaw/acpx#85, merged, pending release).

Tests: 26/26 pass (runtime + tool schema)
Verified e2e: Discord → sessions_spawn(resumeSessionId) → Codex
resumed session and recalled stored secret.

🤖 AI-assisted

* fix: guard resumeSessionId against non-ACP runtime

Add early-return error when resumeSessionId is passed without
runtime="acp" (mirrors existing streamTo guard). Without this,
the parameter is silently ignored and the agent gets a fresh
session instead of resuming.

Also update schema description to note the runtime=acp requirement.

Addresses Greptile review feedback.

* ACP: add changelog entry for session resume (#41847) (thanks @pejmanjohn)

---------

Co-authored-by: Pejman Pour-Moezzi <481729+pejmanjohn@users.noreply.github.com>
Co-authored-by: Onur <onur@textcortex.com>
2026-03-10 10:36:13 +01:00