fc6d4b4010
core: Add User-Agent header to identify client version in HTTP requests
2026-04-25 23:51:23 -04:00
1a734adb4d
core: consolidate shared infrastructure into core package
...
Moves effect logging, observability, runtime utilities, flags, installation
version info, and process utilities from opencode to core package. This
enables better code sharing across packages and establishes core as the
single source of truth for foundational utilities.
All internal imports updated to use @opencode-ai/core paths for consistency.
2026-04-25 13:30:37 -04:00
23a2d01282
fix(observability): standardize session telemetry attrs ( #23213 )
2026-04-17 21:14:23 -04:00
51d8219c46
refactor: unwrap session/ tier-2 namespaces + self-reexport ( #22973 )
2026-04-17 00:49:39 +00:00
675a46e23e
CLI perf: reduce deps ( #22652 )
2026-04-16 02:03:03 -04:00
6b20838981
feat: unwrap provider namespaces to flat exports + barrel ( #22760 )
2026-04-16 05:02:50 +00:00
9f4b73b6a3
fix: clean up final 16 no-unused-vars warnings ( #22751 )
2026-04-16 03:54:21 +00:00
343a564183
feat: unwrap 11 util namespaces to flat exports + barrel ( #22739 )
2026-04-15 23:15:58 -04:00
379e40d772
feat: unwrap InstanceState + EffectBridge namespaces to flat exports + barrel ( #22721 )
2026-04-16 02:45:45 +00:00
1d81335ab5
feat: unwrap Provider namespace + improved automation script ( #22690 )
2026-04-15 21:44:46 -04:00
bbdbc107ae
feat: unwrap Config namespace to flat exports + barrel ( #22689 )
2026-04-15 21:26:24 -04:00
f1751401aa
fix(effect): add effect bridge for callback contexts ( #22504 )
2026-04-15 15:22:34 +00:00
f73ff781e7
fix(opencode): export AI SDK telemetry spans ( #22526 )
2026-04-14 22:30:50 -04:00
f9d99f044d
fix(session): keep GitHub Copilot compaction requests valid ( #22371 )
2026-04-15 08:02:27 +10:00
e8471256f2
refactor(session): move llm stream into layer ( #22358 )
2026-04-13 19:53:30 -04:00
d199648aeb
refactor(permission): remove async facade exports ( #22342 )
2026-04-13 19:33:58 -04:00
c22e34853d
refactor(auth): remove async auth facade exports ( #22306 )
2026-04-13 12:31:43 -04:00
9581bf0670
refactor(effect): upgrade opencode to beta.46 context APIs ( #21977 )
2026-04-10 23:06:28 -04:00
c29392d085
fix: preserve interrupted bash output in tool results ( #21598 )
2026-04-09 10:03:26 -04:00
cd8e8a9928
feat(llm): integrate GitLab DWS tool approval with permission system ( #19955 )
...
Co-authored-by: Aiden Cline <63023139+rekram1-node@users.noreply.github.com >
2026-04-08 20:39:33 -05:00
1f94c48bdd
fix(opencode): keep user message variants scoped to model ( #21332 )
2026-04-07 10:12:53 -04:00
48c1b6b338
tweak: move the max token exclusions to plugins @rekram1-node ( #21225 )
2026-04-06 17:43:58 -05:00
40e4cd27a1
tweak: adjust chat.params hook to allow altering of the maxOutputTokens ( #21220 )
2026-04-06 18:13:30 +00:00
e89527c9f0
feat: Send x-session-affinity and x-parent-session-id headers ( #20744 )
2026-04-02 22:09:53 -05:00
510a1e8140
ignore: fix typecheck in dev ( #20702 )
2026-04-02 15:38:30 +00:00
23c8656080
refactor: split up models.dev and config model definitions to prevent coupling ( #20605 )
2026-04-02 10:00:43 -05:00
567a91191a
refactor(session): simplify LLM stream by replacing queue with fromAsyncIterable ( #20324 )
2026-03-31 15:27:51 -04:00
c5442d418d
refactor(session): effectify SessionPrompt service ( #19483 )
2026-03-30 20:06:51 +00:00
9f3c2bd861
fix: agent value passed to chat.params and chat.headers hooks was not a string ( #19996 )
2026-03-30 16:32:27 +00:00
196a03caff
fix: discourage _noop tool call during LiteLLM compaction ( #18539 )
2026-03-29 22:48:17 -05:00
36aeb32159
ignore: kill todo ( #19566 )
2026-03-29 00:06:32 +00:00
860531c275
refactor(session): effectify session processor ( #19485 )
2026-03-28 12:09:47 -04:00
c33d9996f0
feat: AI SDK v6 support ( #18433 )
2026-03-27 15:24:30 -05:00
9330bc5339
fix: route GitLab Duo Workflow system prompt via flowConfig ( #18928 )
2026-03-24 12:33:18 -05:00
691644eeeb
tweak: add back setting user agent in requests ( #18795 )
2026-03-23 15:34:59 +00:00
38e0dc9ccd
Move service state into InstanceState, flatten service facades ( #18483 )
2026-03-21 04:51:35 +00:00
214a6c6cf1
fix: switch consumers to service imports to break bundle cycles ( #18438 )
2026-03-20 20:55:46 +00:00
05d3e65f76
feat: enable GitLab Agent Platform with workflow model discovery ( #18014 )
2026-03-20 12:55:22 -05:00
bfdc38e421
tweak: adjust codex plugin logic so that codex instruction isn't always added (oauth plan no longer enforces instruction whitelisting) ( #18337 )
2026-03-20 10:37:47 -05:00
1ac1a0287c
anthropic legal requests ( #18186 )
2026-03-19 04:45:24 +00:00
5dfe86dcb1
refactor(truncation): effectify TruncateService, delete Scheduler ( #17957 )
2026-03-17 21:59:54 -04:00
c2ca1494e5
fix(opencode): preserve prompt tool enables with empty agent permissions ( #17064 )
...
Co-authored-by: jquense <jquense@ramp.com >
2026-03-15 23:01:46 -05:00
01d518708a
remove unnecessary deep clones from session loop and LLM stream ( #14354 )
2026-02-19 18:37:55 -06:00
e269788a8f
feat: support claude agent SDK-style structured outputs in the OpenCode SDK ( #8161 )
...
Co-authored-by: Claude Opus 4.5 <noreply@anthropic.com >
Co-authored-by: Dax Raad <d@ironbay.co >
2026-02-12 04:54:05 +00:00
0fd6f365be
fix(core): ensure compaction is more reliable, add reserve token buffer to ensure that input window has enough room to compact ( #12924 )
...
Co-authored-by: James Lal <james@littlebearlabs.io >
2026-02-10 19:55:22 -06:00
ec720145fa
fix: when using codex sub, send the custom agent prompts as a separate developer message (previously sent as user message but api allows for instructions AND developer messages) ( #11667 )
...
Co-authored-by: Carlos <carloscanas942@gmail.com >
2026-02-01 14:57:47 -06:00
d9f18e4006
feat(opencode): add copilot specific provider to properly handle copilot reasoning tokens ( #8900 )
...
Co-authored-by: Claude Opus 4.5 <noreply@anthropic.com >
Co-authored-by: Aiden Cline <63023139+rekram1-node@users.noreply.github.com >
Co-authored-by: Aiden Cline <aidenpcline@gmail.com >
2026-01-30 19:53:22 -06:00
11d486707c
fix: rm ai sdk middleware that was preventing <think> blocks from being sent back as assistant message content ( #11270 )
...
Co-authored-by: opencode-agent[bot] <opencode-agent[bot]@users.noreply.github.com>
2026-01-30 09:36:57 -05:00
03ba49af4e
fix(telemetry): restore userId and sessionId metadata in experimental_telemetry ( #8195 )
2026-01-29 10:14:28 -06:00
29ea9fcf25
fix: ensure variants for copilot models work w/ maxTokens being set
2026-01-28 21:55:50 -06:00
Dax Raad