Patrick Erichsen
efb7d426cf
fix(browser): reject ax<N> refs in act path instead of timing out ( #69924 )
2026-04-21 18:43:27 -07:00
claycurry
6a68f1dd57
Docs: link feature cards to relevant pages
...
Link docs feature cards to their intended destination pages in the English docs surfaces.
- add hrefs to the feature cards in docs/concepts/features.md
- add hrefs to the key capability cards in docs/index.md
- preserve current main branch copy while landing the navigation fix
2026-04-21 20:36:55 -05:00
Peter Steinberger
fb9a21ae8f
fix: centralize draft preview finalization
2026-04-22 02:32:55 +01:00
Peter Steinberger
ffef84dea7
ci: start runtime tests without dist
2026-04-22 02:27:03 +01:00
Peter Steinberger
e5909f3e5d
ci: scope mlx helper as macos native
2026-04-22 02:19:58 +01:00
Peter Steinberger
e836b5b6d7
ci: isolate mlx from macos swift checks
2026-04-22 02:12:07 +01:00
Peter Steinberger
710e4e9e51
ci: widen package boundary cache inputs
2026-04-22 01:53:22 +01:00
Gustavo Madeira Santana
f4478a142a
Fix channel presence gating for disabled plugins ( #69862 )
...
Merged via squash.
Prepared head SHA: f76f6212b25599352+gumadeiras@users.noreply.github.com >
Co-authored-by: gumadeiras <5599352+gumadeiras@users.noreply.github.com >
Reviewed-by: @gumadeiras
2026-04-21 20:51:09 -04:00
Peter Steinberger
eb6006730d
fix(line): guard outbound media targets
2026-04-22 01:48:14 +01:00
Peter Steinberger
66576f3355
test(extensions): fix lint-clean test assertions
2026-04-22 01:43:18 +01:00
Peter Steinberger
d57fe63ee0
ci: cache package boundary artifacts
2026-04-22 01:42:44 +01:00
Peter Steinberger
5c74e9da01
fix(qqbot): avoid eager storage directory creation
2026-04-22 01:42:10 +01:00
Peter Steinberger
540171ddbd
docs: clarify ACP delivery model
2026-04-22 01:32:20 +01:00
Peter Steinberger
73d9746e6a
ci: reuse swift build cache for unchanged inputs
2026-04-22 01:30:40 +01:00
Peter Steinberger
ce05418930
ci: preserve exact swift build cache
2026-04-22 01:26:05 +01:00
Gustavo Madeira Santana
819d15481d
fix: validate plugin source entries before runtime inference ( #69868 )
...
Merged via squash.
Prepared head SHA: b67644cdda5599352+gumadeiras@users.noreply.github.com >
Co-authored-by: gumadeiras <5599352+gumadeiras@users.noreply.github.com >
Reviewed-by: @gumadeiras
2026-04-21 20:12:19 -04:00
Gustavo Madeira Santana
19354c9a6a
fix(discord): keep slash follow-ups ephemeral ( #69869 )
...
Merged via squash.
Prepared head SHA: 0f5ab771565599352+gumadeiras@users.noreply.github.com >
Co-authored-by: gumadeiras <5599352+gumadeiras@users.noreply.github.com >
Reviewed-by: @gumadeiras
2026-04-21 20:02:59 -04:00
Ron Cohen
08bc16853e
WhatsApp: add group and direct system prompt support ( #59553 )
...
Merged via squash.
Prepared head SHA: 63e2b50e011525690+Bluetegu@users.noreply.github.com >
Co-authored-by: omarshahine <10343873+omarshahine@users.noreply.github.com >
Reviewed-by: @omarshahine
2026-04-21 16:40:32 -07:00
Gustavo Madeira Santana
06a6dd5a6b
chore(docs): update stale docs ref
2026-04-21 19:33:36 -04:00
Peter Steinberger
37463af5e1
ci: increase package boundary compile concurrency
2026-04-22 00:26:03 +01:00
Onur Solmaz
99787dbf45
docs(skills): add duplicate triage maintainer skill ( #69780 )
...
* docs(skills): add duplicate triage maintainer skill
* docs(skills): align duplicate triage with prtags sync
* docs(skills): remove local paths from duplicate triage skill
* docs(skills): use pr-search-cli naming consistently
* docs(skills): fix pr-search-cli command usage
* docs(skills): use tested release install commands
* docs(skills): treat prtags comment sync as automatic
* docs(skills): adjust duplicate triage skill title
* docs(skills): add duplicate triage UI metadata
2026-04-22 01:18:07 +02:00
Peter Steinberger
85c63942a5
ci: skip swift package patch in macos node lane
2026-04-22 00:16:45 +01:00
Peter Steinberger
a426ef5b6a
ci: preserve swift build cache hits
2026-04-22 00:12:03 +01:00
Bek
e116b343b2
feat(slack): Annotate inbound Slack mention tokens in Slack RawBody and BodyForAgent content so the agent sees both the actionable Slack mention token and a human-readable name. ( #65731 )
...
* Annotate inbound Slack mentions in raw bodies
* Avoid shared regex state in Slack mention rendering
* Bound Slack mention lookups with concurrency
* slack: keep mention concurrency helper plugin-local
* test: stabilize node core CI assertions
* slack: cap mention lookups per inbound message
* test: reset suite gateway runtime state
* fix(slack): reuse plugin sdk concurrency helper
2026-04-21 19:03:50 -04:00
Peter Steinberger
6bf56d8637
ci: cap android checkout and use build cache
2026-04-22 00:02:40 +01:00
Peter Steinberger
cc8ecde364
ci: avoid external gradle action in android checks
2026-04-21 23:56:52 +01:00
Peter Steinberger
6966f018f7
ci: quiet mlx swift manifest warnings
2026-04-21 23:52:04 +01:00
Peter Steinberger
e822e71410
ci: cap stuck checkout retries
2026-04-21 23:47:17 +01:00
Peter Steinberger
df3fcbd716
test: lazy-load openai provider catalog contract
2026-04-21 23:35:37 +01:00
Bek
70683179a0
fix(slack): narrow first turn context seeding to remove redundant thread-starter content ( #68402 )
...
Fix Slack thread bootstrap replaying the bot's own prior turns into new sessions and duplicating the thread-starter prompt block.
Narrows first-turn context seeding to exclude only the current Slack bot's own starter/history entries, so self-authored turns no longer pollute new session prompts while preserving human and third-party bot context
Removes the redundant plain-text starter prelude in runPreparedReply() that doubled thread-starter content when no ThreadHistoryBody was present
2026-04-21 18:28:34 -04:00
Peter Steinberger
acf67c1a42
docs: tighten optimizetests skill
2026-04-21 23:24:51 +01:00
Bek
dfe0e49c8a
fix(qmd): Dedup in-flight manager creation so only one full QMD manager arms per agent/config at a time, eliminating the concurrent exportSessions() collisions that triggered path changed during write errors ( #65226 )
...
Fixes concurrent manager creation races that caused SafeOpenErrors during session export.
Deduplicates in-flight manager creation so only one full QMD manager arms per agent/config at a time, eliminating the concurrent exportSessions() collisions that triggered path changed during write errors
Resolves and snapshots runtime inputs before cache reuse, replacing stale managers atomically when workspace/config changes, and aborting queued export work promptly on close()
2026-04-21 18:22:21 -04:00
Bek
1acb094579
fix: wrap oversized session lines before JSONL write ( #64494 )
...
updates the real session-export path so pathological transcript messages no longer become a single toxic export line for downstream indexing.
2026-04-21 18:18:22 -04:00
Gustavo Madeira Santana
66add9fcd9
perf(cli): lazy-load doctor plugin paths ( #69840 )
...
Merged via squash.
Prepared head SHA: ebf93ad9135599352+gumadeiras@users.noreply.github.com >
Co-authored-by: gumadeiras <5599352+gumadeiras@users.noreply.github.com >
Reviewed-by: @gumadeiras
2026-04-21 18:17:19 -04:00
Bek
0e1d324dd8
fix(agents): Wake active requester sessions for subagent completions while keeping dormant sessions externally deliverable ( #62963 )
...
Route subagent completion announces through embedded-run wake for active requesters, preserve external delivery for dormant ones
2026-04-21 18:13:53 -04:00
Bek
14dcbd4044
fix(prompt): align system prompt messaging and subagent routing guidance ( #64059 )
...
replace legacy `to` with `target` in prompt
2026-04-21 18:10:53 -04:00
Peter Steinberger
824c4785e4
test: speed channel contract suites
2026-04-21 23:09:22 +01:00
Devin Robison
ee316dbc4b
fix(tlon): guard memex upload target ( #69794 )
...
* fix(tlon): guard memex upload target
* fix(tlon): harden guarded memex upload
* fix(tlon): validate hosted memex upload targets
* fix(tlon): tighten hosted domain matching
* fix(tlon): reject non-standard memex upload ports
* fix(tlon): disable memex upload redirects
* test(tlon): drop redundant mock resets in memex upload test
* chore(lint): update tlon raw-fetch allowlist for guarded memex upload
* fix(tlon): reject unparseable ship URLs in hosted-ship classifier
* fix(lint): point tlon raw-fetch allowlist at fetch callee lines
* fix(tlon): guard custom-S3 upload through fetchWithSsrFGuard
* fix(tlon): preserve scheme-less hosted ship routing and allow explicit :443
* docs(changelog): note tlon upload guard
* fix(tlon): guard memex lookup and private s3 opt-in
* fix(tlon): validate upload result URLs
2026-04-21 15:57:49 -06:00
Peter Steinberger
74668ea8a1
fix(image-generation): log provider fallback failures
2026-04-21 22:50:09 +01:00
Vincent Koc
b5c4aaf2a7
fix(install): mirror node-domexception override for npm ( #69819 )
...
* fix(install): mirror node-domexception override for npm
* docs(changelog): credit npm install override fix
* fix(install): pin domexception override exactly
* docs(changelog): drop leftover npm fix merge markers
* Update CHANGELOG.md
2026-04-21 14:45:05 -07:00
Peter Steinberger
d1e3789e15
test: optimize slow test hotspots
2026-04-21 22:42:08 +01:00
Bek
49b233caa1
fix(slack): preserve thread aliases in runtime outbound sends ( #62947 )
...
Slack-threaded direct sends that go through the generic runtime wrapper now stay in the intended thread when the caller supplies threadTs.
2026-04-21 17:40:47 -04:00
Vincent Koc
475e6ff1d1
docs(gateway): replace user-facing 'extension' references with 'plugin' per terminology rules
2026-04-21 14:39:10 -07:00
Peter Steinberger
d2f68af615
docs: document Ollama image understanding
2026-04-21 22:33:56 +01:00
Vincent Koc
f1f6214fd5
docs(help): add frontmatter to gpt54-codex parity docs
2026-04-21 14:29:58 -07:00
Peter Steinberger
e71e543350
fix: route explicit image describe models
2026-04-21 22:25:45 +01:00
Peter Steinberger
a7ff7dd945
docs: note Ollama image routing ( #69816 ) (thanks @soloclz)
2026-04-21 22:25:45 +01:00
soloclz
9a22cd212b
fix(ollama): register media-understanding provider so image tool can route ollama/* models
...
Ollama chat models already support image inputs (extensions/ollama/src/stream.ts
extracts image parts and forwards them via the Ollama API), but the ollama
plugin did not register a MediaUnderstandingProvider. The image tool's provider
registry therefore had no 'ollama' entry, so requests like
`imageModel: 'ollama/qwen2.5vl:7b'` failed to resolve and fell back to
unrelated providers.
Register ollamaMediaUnderstandingProvider with:
- capabilities: ['image']
- describeImage/describeImages wired to the shared core helpers (reuses the
same pi-ai complete path Ollama chat already goes through)
- no defaultModels or autoPriority: Ollama vision support depends on which
model the user has pulled, so we don't pick a canonical default and don't
auto-steal image duty from configured providers.
Fixes #69071 (and supersedes #60280 ).
2026-04-21 22:25:45 +01:00
Vincent Koc
b2f96f7f05
docs(providers): alphabetize Cloudflare/ComfyUI and vLLM/Vydra entries
2026-04-21 14:25:31 -07:00
Devin Robison
7be82d4fd1
fix(openshell): pin host writes to sandbox root ( #69797 )
...
* fix(openshell): pin host writes to sandbox root
* fix(openshell): use plugin sdk infra runtime
* fix(openshell): reject symlink write targets
* chore(changelog): note openshell sandbox write fix
2026-04-21 15:18:28 -06:00
