open-multi-agent-kit 1.2.0 → 1.2.3

package/docs/runtime-algorithms.md

@@ -21,6 +21,9 @@ baseline below is history, not current truth.
 | Provider retry/failover classification | yes | `provider-retry`, `session-failure-cause` | default | resilience/classification tests | released | not measured |
 | MCP descriptor injection screen | yes | `mcp/manager` import path | default | quarantine tests | released | pattern rule score, not calibrated risk |
 | verified-run coordinator + evidence | yes | verified-run paths | opt-in command | coordinator/evidence tests | working tree | scope-limited binding, not general correctness |
+| Atomic commit planner (`planAtomicCommits`) | yes | public agent API only; no live commit caller | explicit function call | local planner/API/property tests, not a CI receipt | working tree | not measured |
+
+Named tests are evidence locations, not a claim that this exact working tree passed remote CI. Source call paths and fresh local results must be checked before promoting a gate.
 
 Gates are reported per row so a green "implemented" never upgrades itself to
 "released" or "measured".
@@ -61,6 +64,45 @@ and the authorization hook still runs once. Coverage:
 boundary only; isolating a non-cooperative extension needs a killable execution
 boundary, which this change does not add.
 
+## Deferred claim freshness and cancelled memo requests (2026-09-22)
+
+The live path is `agentLoop` → `schedulePlannedDagLevels` → `runDagFrontier`.
+A deferred call retains its prepared arguments and one-time authorization, and
+admission always re-resolves its claims: a cached resolution that survived the
+wait could execute a stale scope (A holds X, B defers on X, C starts on Z, and
+after A settles B runs on Z using the cached X claim). Re-resolution preserves
+conflict exclusion against both earlier pending calls and later running calls,
+and stays abort-bound, so a late claim result cannot admit an aborted call.
+
+The cached resolution is still reused for one thing — proving the deferral.
+A deferral decision is only "wait", so a stale conflict answer executes nothing,
+while re-invoking an extension `resourceClaims()` callback on every scan of the
+wait paid real I/O for the same answer. The resolver-budget regression shows a
+deferred call blocked behind a slow peer invoking the callback 5 times when the
+answer was re-resolved per scan, and 3 times with the split: the scheduling
+resolution, one conflict-proving resolution, and the fresh resolution required
+before admission. Claim callbacks may still run again after a wait, so they must
+not be used as authorization side effects.
+
+The per-run schedule memo checks cancellation before inspecting arguments or
+serializing its key, including warm-cache requests. An already-aborted request
+returns no plan and does not promote cache recency.
+
+`assignDagDependencies` also returns an empty-edge graph directly when every
+resolution is read-only (including empty claims). In the 128-entry regression,
+resolved-entry visits fell from 16,256 to 128 with the same dependency graph.
+Mixed writes and exclusive barriers still use the existing conflict scan. This
+is an operation-count result for that fixture, not measured end-to-end latency,
+model quality, CI certification or a release claim.
+
+Coverage: `tool-dag-deferred-refresh.test.ts`,
+`tool-dag-deferred-resolver-budget.test.ts`,
+`tool-dag-memo-cancellation.test.ts`, and
+`tool-dag-readonly-dependencies.test.ts` in `packages/agent/test/`.
+The separate `planAtomicCommits` export remains a pure public API,
+not a live automatic commit coordinator. ECRAF and automatic shards
+are not activated by these changes.
+
 ## Reasoning router resolver contract (2026-09-19 audit F05/F06)
 
 The low-confidence escalation in `resolveThinkingLevelV4WithUncertainty` is
@@ -173,16 +215,19 @@ and places conflicting calls in later levels. Before a level executes, OMK autho
 re-plans with post-hook arguments so a hook cannot silently invalidate the
 original claim plan.
 
-The live executor uses level barriers. `assignDagDependencies()` computes a
-finer predecessor graph, but no live executor consumes it. Each candidate level
-is authorized and then re-planned from post-hook arguments.
+The live executor uses a ready frontier, not level barriers.
+`assignDagDependencies()` computes the predecessor graph and the loop feeds it
+to `runDagFrontier()`, which admits a node as soon as its predecessors settle
+and a resident slot frees up. Each candidate is still authorized and
+re-planned from post-hook arguments before execution.
 
 Evidence:
 
 - `packages/agent/src/tool-dag-scheduler.ts`: `assignDagLevels`,
   `assignDagDependencies`, `scheduleDagLevels`
 - `packages/agent/src/agent-loop.ts`: `executeToolCallsDagLevels`,
-  `runDagLevelCalls`
+  `runDagFrontier` (predecessor-settle ready queue; the former
+  `runDagLevelCalls` level barrier was replaced)
 - `packages/agent/test/tool-dag-scheduler*.test.ts`
 - `packages/agent/test/tool-dag-dependencies.test.ts`
 

package/docs/session-format.md

@@ -372,6 +372,14 @@ Key methods for working with sessions programmatically.
 ### Static Listing Methods
 - `SessionManager.list(cwd, sessionDir?, onProgress?)` - List sessions for a directory
 - `SessionManager.listAll(onProgress?)` - List all sessions across all projects
+- `SessionManager.list(cwd, sessionDir, onProgress, { metadataOnly: true })` - Return metadata without `allMessagesText`
+- `SessionManager.listAll(sessionDir, onProgress, { metadataOnly: true })` - Metadata-only listing; pass `undefined` for global scope
+
+Existing calls retain `SessionInfo[]` and its full `allMessagesText`. The opt-in
+metadata calls return the same metadata fields without that property; absence
+means not loaded, not empty history. The TUI uses this opt-in and performs bounded
+full-text reads on search. There is no disk format or sidecar migration, and no
+change to append, complete-entry parsing, branch resolution, or export.
 
 ### Instance Methods - Session Management
 - `newSession(options?)` - Start a new session (options: `{ parentSession?: string }`)

package/docs/sessions.md

@@ -69,6 +69,25 @@ In the picker you can:
 
 When available, omk uses the `trash` CLI for deletion instead of permanently removing files.
 
+### Listing and full-history search
+
+The `/resume` and `--resume` pickers keep metadata (name, first message, dates,
+count and ancestry), not every session's concatenated message text. Typing a
+query reads full user/assistant text from the original JSONL files, at most two
+files at a time. Fuzzy tokens, quoted phrases, regex, ordering, named filters and
+all-project scope remain available. Search covers all stored message branches,
+not only visible rows or the current branch.
+
+A loading state is shown until search settles. Changing the query or scope
+cancels obsolete work; replacement work waits for its reads to settle, and old
+results cannot overwrite the current query. Missing/unreadable or replaced
+session identities are reported rather than treated as fully searched. Each
+query rereads the source, so appended text is searchable without a stale sidecar.
+
+This does not change JSONL durability, resume, copying, export, or branch storage.
+No disk index is introduced. A search still temporarily materializes up to two
+full text projections; one very large entry or first message can still be large.
+
 ## Naming Sessions
 
 Use `/name <name>` to set a human-readable session name:

package/docs/skills.md

@@ -77,6 +77,25 @@ For project-level Claude Code skills, add to `.omk/settings.json`:
 
 This is progressive disclosure: only descriptions are always in context, full instructions load on-demand.
 
+### Catalog cache
+
+Startup caches skill metadata in `cache/skill-catalog-v2.json` under the agent
+directory. Unchanged, complete fingerprints allow reuse; `.gitignore`, `.ignore`,
+and `.fdignore` changes invalidate discovery just like skill file edits. A depth
+or entry limit, unreadable path, or symlink cycle makes the fingerprint incomplete,
+so OMK uses the original scanner rather than a partial cached inventory. Results
+are stored only when fingerprints agree before and after a scan.
+
+Writers use exclusive, owner-only temporary files and atomic rename. Concurrent
+writers may replace each other's cache snapshots (causing misses), but do not share
+or remove each other's temporary files. The 64-entry limit bounds cached roots,
+not the number of skills. Cache hits return cloned metadata; corrupt entries and
+legacy v1 files are misses. Legacy files are not deleted or migrated in place.
+
+This is metadata-based invalidation, not an atomic filesystem snapshot, and does
+not change scanner precedence, trust rules, or skill availability. Cold scans pay
+for a second fingerprint walk; incomplete trees cannot use the cache fast path.
+
 ### Default active skills
 
 Set `defaultActiveSkills` in global settings to mark selected discovered skills active in every prompt without repeating `!skill:` invocations:

package/docs/terminal-setup.md

@@ -69,6 +69,21 @@ Add to `keybindings.json` to enable `Shift+Enter` for multi-line input:
 }
 ```
 
+## Zed (Integrated Terminal)
+
+Add these key bindings to your Zed `keymap.json`:
+
+```json
+{
+  "context": "Terminal",
+  "bindings": {
+    "shift-enter": ["terminal::SendText", "\u001b[13;2u"],
+    "ctrl--": ["terminal::SendText", "\u001b[45;5u"],
+    "ctrl-alt-]": ["terminal::SendText", "\u001b[93;7u"]
+  }
+}
+```
+
 ## Windows Terminal
 
 Add to `settings.json` (Ctrl+Shift+, or Settings → Open JSON file) to forward the modified Enter keys omk uses:

package/docs/tui.md

@@ -452,6 +452,21 @@ interface MyTheme {
 }
 ```
 
+## Output observations
+
+`ProcessTerminal.getOutputStats()` returns numeric counters for write calls,
+UTF-8 bytes offered, `write(false)`, drain/error events, and current/peak writable
+buffer length. `/debug` displays them and `/debug save` includes only these
+allowlisted fields. No prompt, image, ANSI frame, or error message is retained in
+the counters. Terminals without this optional API omit the observations.
+
+Every ProcessTerminal paint, cursor, title, progress, protocol and restoration
+write goes through the observer. A false return is counted, never retried: the
+Writable has already accepted the data. Error monitoring does not consume normal
+error handling; listeners are detached on stop. This is instrumentation, not a
+backpressure scheduler or an output-memory bound. Intermediate paint coalescing
+and bounded shutdown flushing remain separate work.
+
 ## Debug logging
 
 Set `OMK_TUI_WRITE_LOG` to capture the raw ANSI stream written to stdout.
@@ -488,6 +503,32 @@ class CachedComponent {
 
 Call `invalidate()` when state changes, then `handle.requestRender()` to trigger re-render.
 
+### Assistant streaming blocks
+
+`AssistantMessageComponent` reuses text/thinking components at unchanged content
+indices. It updates a block only when its displayed text changes, preserving the
+real Markdown render cache for stable blocks even when the message is mutated in
+place. Removed blocks are released; thinking visibility changes replace the affected
+representation. Width changes reflow Markdown, and explicit invalidation rebuilds
+views for theme changes, including pre-styled hidden-thinking labels.
+
+This does not coalesce model events, truncate history, or incrementally parse an
+actively changing Markdown block. Error/abort text, spacing, and OSC 133 markers
+retain their existing behavior.
+
+### Tool image conversion ownership
+
+Kitty-only non-PNG conversions are shared by source digest and MIME type, not
+image position. Repeated results and reordered images reuse active work; late
+results cannot replace a newer source. Hidden images defer new conversion work,
+PNG is passed through, and original tool/session image data is never deleted.
+
+Conversion failure leaves a visible fallback and is retried on an explicit result
+update, not on every render. Clearing chat history or stopping interactive mode
+disposes its image projections and prevents late repaints. The converter has no
+cancellation API: in-flight work remains owned until its promise settles.
+Only derived images referenced by the current result are cached.
+
 ## Invalidation and Theme Changes
 
 When the theme changes, the TUI calls `invalidate()` on all components to clear their caches. Components must properly implement `invalidate()` to ensure theme changes take effect.

package/docs/verified-run-safety.md

@@ -0,0 +1,75 @@
+# Verified run의 안전 경계
+
+`linux-command-v1`, `linux-command-dag-v1`, synthetic `linux-scripted-agent-v1`의 소스 구현을 설명합니다. 실제 provider 연결, 자동 shard 실행, active ECRAF, 원격 게시, managed apply의 지원 선언은 아닙니다.
+
+## 권한과 복구
+
+- `AuthorityStore`는 실행 시작 시 host clock을 읽습니다. 시각 인자를 생략해도 만료 검사를 수행하며, deadline과 같은 시각부터 신규 시작을 거절합니다. 테스트용 clock은 trusted host 구성에서 주입합니다.
+- 현재 epoch, incarnation, token, claims와 실행 상태를 확인합니다. `lookup`의 `pending`은 조회 결과이며 실행 승인이 아닙니다.
+- 게시 재시도는 기존 ref를 먼저 관찰합니다. CAS 결과만 기록하는 복구에는 새 effect를 시작하지 않습니다. `running`, `quarantined`, terminal grant를 새로운 CAS에 재사용하지 않습니다.
+- 취소나 만료는 종료 증명이 아닙니다. 실행 identity가 없는 crash 기록은 계속 `unknown`으로 남습니다.
+- dispatch journal 쓰기가 spawn 전에 실패하면 reservation을 정산합니다. `onReady` 이후의 오류를 pre-spawn 오류로 오인해 권한을 해제하지 않습니다.
+
+## 후보와 Git
+
+전체 manifest의 경로, 부모 디렉터리, 중복, 파일/디렉터리 충돌, POSIX mode, 크기, 내용 digest를 첫 object write 전에 확인합니다. Git으로 표현할 수 없는 mode는 거절하며, 검증 뒤 조용히 정규화하지 않습니다. `mktree -z`로 이름을 보존하고 sealed tree를 manifest와 다시 비교합니다. SHA-1과 SHA-256 저장소를 처리합니다.
+
+Git 호출은 무작위 private hooks 디렉터리를 만들고 owner, 0700 mode, symlink 여부, 비어 있는지를 확인합니다. child 종료 뒤 자기 디렉터리만 정리합니다. 상속 Git config, hooks, fsmonitor, 자동 maintenance와 signing에 의존하지 않습니다. 같은 UID 또는 host 관리자 침해를 방어한다는 주장은 하지 않습니다.
+
+현재 inclusion은 `.git`과 `.omk`를 제외한 전체 트리이며 알려진 민감 파일 이름을 거절합니다. 저장된 legacy 후보도 게시 전에 재검사합니다. 이는 비밀 내용 탐지기가 아니며, 사용자별 inclusion policy digest를 contract와 receipt에 연결하는 확장은 별도 작업입니다.
+
+## 게시 deadline과 취소
+
+`RunCoordinator.publish`와 CLI publish는 같은 취소 신호를 사용합니다. Git 객체 생성 전과 CAS 전에 신호와 권한을 다시 확인합니다. CAS 성공 뒤의 취소는 ref를 되돌리지 않습니다.
+
+공개 SDK/CLI의 `PublishOptions.timeoutMs`는 1~60,000ms이며 기본값은 60초입니다. host와 worker는 동일한 Linux monotonic deadline을 사용합니다. worker의 각 Git child는 남은 예산과 30초 중 작은 timeout을 사용합니다. 같은 blob digest는 한 번만 기록합니다.
+
+### Owned Git publisher
+
+`RunCoordinator.publish`와 CLI publish는 `owned-git-v1`을 사용합니다. 검증용 namespace 감독자를 재사용하며 worktree는 read-only, 실제 `.git` 디렉터리만 writable로 mount합니다. 현재 Node 실행 파일과 고정된 worker 모듈 집합을 사용하고, 사용자 extension이나 모델이 지정한 모듈은 로드하지 않습니다. 모듈과 실행 파일 digest는 실행 의미에 결합합니다.
+
+순서는 durable dispatch intent → namespace identity 기록 → 후보 sealing → publish intent 기록 → fresh authority 확인 → candidate-OID에 결합된 CAS gate → ref 관측 → child close와 namespace drain → 권한 정산입니다. PID 숫자만으로 종료를 인정하지 않습니다. PID birth 또는 boot identity가 다르면 `unknown`이며, 실제 namespace가 비었을 때만 종료를 확인합니다.
+
+CAS 결과와 process 종료는 별도 사실입니다. ref가 이미 candidate라면 다시 CAS하지 않고 metadata만 복구할 수 있습니다. Git grant가 아직 남아 있으면 status는 `pending_effects`, `cleanSuccess: false`, `terminal: false`를 표시합니다. 취소가 CAS 이후 도착하면 이미 반영된 ref를 되돌리지 않습니다. callback 실패에도 child를 abort하고 종료를 관측하며, 증명이 부족하면 grant와 private staging을 보존합니다.
+
+초기 지원 대상은 Linux+bwrap+Node의 source/npm 실행이며 `.git`이 실제 디렉터리인 저장소입니다. linked worktree의 외부 git directory, Bun/standalone bundle의 worker 파일 미제공, namespace 사용 불가에는 명시적으로 실패합니다. 동기 helper `publishVerifiedRun`은 내부 호환 경로로 남지만 SDK/CLI가 그 경로로 자동 후퇴하지 않습니다.
+
+벽시계 역행은 `clock_anomaly`로 차단합니다. 이 오류를 없애려고 journal을 지우거나 시계 검사를 끄지 마십시오. workspace/fsync 같은 host 동기 I/O 전체를 OS 수준에서 시간 제한한다는 보장은 하지 않습니다.
+
+## 증거 조회
+
+SDK의 `evidenceRead(runId)`와 `omk run evidence ID --json`은 다음 축을 구분합니다.
+
+| 필드 | 의미 |
+| --- | --- |
+| `authenticity` | `valid`, `invalid`, `unverifiable` |
+| `verification` | `passed`, `failed`, `incomplete` |
+| `currentEnvironmentEligibility` | `matching`, `different`, `unsupported`, `unknown` |
+
+정상 서명된 실패 영수증은 `valid`와 `failed`입니다. key가 없으면 `unverifiable`, 변조를 확인하면 `invalid`입니다. 현재 sandbox나 실행 파일이 없어도 과거 증거를 읽을 수 있습니다. 게시와 재실행은 별도의 현재 환경 검사를 유지합니다. 조회는 journal을 수정하거나 영수증을 재서명하지 않습니다.
+
+## 저널의 incremental replay
+
+같은 inode와 head 계보에서 읽은 committed prefix 바이트가 캐시와 정확히 같을 때만 suffix를 파싱합니다. 크기나 timestamp 일치만으로 prefix를 신뢰하지 않습니다. 교체, 축소, prefix 변경, GC, 재시작에서는 full replay 또는 오류로 돌아갑니다.
+
+head CAS, reducer 사전검사, append fsync, head 게시, 재읽기 검증은 그대로 수행합니다. 외부 조회에는 복사된 projection을 반환해 caller가 cached state를 바꾸지 못하게 합니다. `OpenAuthorityStoreOptions.incrementalReplay: false`로 파싱 재사용만 끌 수 있으며 저장 형식은 바뀌지 않습니다. 전체 파일 읽기는 남아 있으므로 O(1) I/O나 일정한 mutation latency를 보장하지 않습니다.
+
+## Shard helper
+
+heavy-process capacity 0은 실행 0입니다. 실행 전 journal 쓰기가 실패하면 획득한 permit을 반환합니다. 예외가 나도 시작한 형제 작업을 abort하고 join합니다. runner 거절만으로 종료를 추정하지 않으며, 불명확한 작업은 permit과 journal 소유권을 유지합니다.
+
+orphan attempt의 재시작은 trusted `observeTermination`이 해당 attempt의 종료를 확인한 뒤에만 가능합니다. `ShardRunner`는 자신이 소유한 실행 경계의 종료를 관찰한 후 resolve해야 합니다. 이 helper를 고쳤다는 사실이 verified DAG나 실제 provider에 자동 연결됐다는 뜻은 아닙니다.
+
+## 회귀 검사
+
+저장소 루트에서 좁은 offline 검사를 실행합니다.
+
+```bash
+LIVE_E2E=0 node node_modules/vitest/dist/cli.js --run \
+  packages/coding-agent/test/safe-wiring \
+  packages/coding-agent/test/verified-run \
+  packages/coding-agent/test/workload-shard --maxWorkers=2
+npm run check
+```
+
+전체 검증은 [Development](development.md)의 빌드 및 키 없는 테스트 절차를 따릅니다. 실제 계정의 인증 파일을 이동하지 않도록 검증용 checkout과 별도 HOME을 사용합니다. local pass, 설치 artifact smoke, 원격 CI, 실모델 효과 측정은 서로 다른 증거입니다.

package/examples/extensions/custom-provider-anthropic/index.ts

@@ -378,7 +378,7 @@ function streamCustomAnthropic(
 					// Anthropic gates newer models on this version (400
 					// claude_code_version_too_old). Canonical value: CLAUDE_CODE_VERSION
 					// in omk-ai (packages/ai/src/utils/claude-code-identity.ts).
-					"user-agent": "claude-cli/2.1.258 (external, cli)",
+					"user-agent": "claude-cli/2.1.280 (external, cli)",
 					"x-app": "cli",
 				};
 			} else {

package/examples/extensions/custom-provider-anthropic/package-lock.json

@@ -1,12 +1,12 @@
 {
   "name": "omk-extension-custom-provider-anthropic",
-  "version": "1.2.0",
+  "version": "1.2.3",
   "lockfileVersion": 3,
   "requires": true,
   "packages": {
     "": {
       "name": "omk-extension-custom-provider-anthropic",
-      "version": "1.2.0",
+      "version": "1.2.3",
       "dependencies": {
         "@anthropic-ai/sdk": "0.52.0"
       }

package/examples/extensions/custom-provider-anthropic/package.json

@@ -1,7 +1,7 @@
 {
 	"name": "omk-extension-custom-provider-anthropic",
 	"private": true,
-	"version": "1.2.0",
+	"version": "1.2.3",
 	"type": "module",
 	"scripts": {
 		"clean": "echo 'nothing to clean'",

package/examples/extensions/custom-provider-gitlab-duo/package.json

@@ -1,7 +1,7 @@
 {
 	"name": "omk-extension-custom-provider-gitlab-duo",
 	"private": true,
-	"version": "1.2.0",
+	"version": "1.2.3",
 	"type": "module",
 	"scripts": {
 		"clean": "echo 'nothing to clean'",

package/examples/extensions/doom-overlay/wad-finder.ts

@@ -1,11 +1,12 @@
 import { existsSync, writeFileSync } from "node:fs";
 import { dirname, join, resolve } from "node:path";
 import { fileURLToPath } from "node:url";
+import { gunzipSync } from "node:zlib";
 
 // Get the bundled WAD path (relative to this module)
 const __dirname = dirname(fileURLToPath(import.meta.url));
 const BUNDLED_WAD = join(__dirname, "doom1.wad");
-const WAD_URL = "https://distro.ibiblio.org/slitaz/sources/packages/d/doom1.wad";
+const WAD_URL = "https://www.gamers.org/pub/idgames/idstuff/doom/doom-1.8.wad.gz";
 
 const DEFAULT_WAD_PATHS = ["./doom1.wad", "./DOOM1.WAD", "~/doom1.wad", "~/.doom/doom1.wad"];
 
@@ -42,8 +43,11 @@ export async function ensureWadFile(): Promise<string | null> {
 		if (!response.ok) {
 			throw new Error(`HTTP ${response.status}`);
 		}
-		const buffer = await response.arrayBuffer();
-		writeFileSync(BUNDLED_WAD, Buffer.from(buffer));
+		const wad = gunzipSync(Buffer.from(await response.arrayBuffer()));
+		if (wad.subarray(0, 4).toString("ascii") !== "IWAD") {
+			throw new Error("Downloaded file is not a valid IWAD");
+		}
+		writeFileSync(BUNDLED_WAD, wad);
 		return BUNDLED_WAD;
 	} catch {
 		return null;

package/examples/extensions/gondolin/package-lock.json

@@ -1,12 +1,12 @@
 {
 	"name": "omk-extension-gondolin",
-	"version": "1.2.0",
+	"version": "1.2.3",
 	"lockfileVersion": 3,
 	"requires": true,
 	"packages": {
 		"": {
 			"name": "omk-extension-gondolin",
-			"version": "1.2.0",
+			"version": "1.2.3",
 			"dependencies": {
 				"@earendil-works/gondolin": "0.12.0"
 			}

package/examples/extensions/gondolin/package.json

@@ -1,7 +1,7 @@
 {
 	"name": "omk-extension-gondolin",
 	"private": true,
-	"version": "1.2.0",
+	"version": "1.2.3",
 	"type": "module",
 	"scripts": {
 		"clean": "echo 'nothing to clean'",

package/examples/extensions/sandbox/package-lock.json

@@ -1,12 +1,12 @@
 {
 	"name": "omk-extension-sandbox",
-	"version": "1.2.0",
+	"version": "1.2.3",
 	"lockfileVersion": 3,
 	"requires": true,
 	"packages": {
 		"": {
 			"name": "omk-extension-sandbox",
-			"version": "1.2.0",
+			"version": "1.2.3",
 			"dependencies": {
 				"@anthropic-ai/sandbox-runtime": "0.0.26"
 			}

package/examples/extensions/sandbox/package.json

@@ -1,7 +1,7 @@
 {
 	"name": "omk-extension-sandbox",
 	"private": true,
-	"version": "1.2.0",
+	"version": "1.2.3",
 	"type": "module",
 	"scripts": {
 		"clean": "echo 'nothing to clean'",

package/examples/extensions/with-deps/package-lock.json

@@ -1,12 +1,12 @@
 {
   "name": "omk-extension-with-deps",
-  "version": "1.2.0",
+  "version": "1.2.3",
   "lockfileVersion": 3,
   "requires": true,
   "packages": {
     "": {
       "name": "omk-extension-with-deps",
-      "version": "1.2.0",
+      "version": "1.2.3",
       "dependencies": {
         "ms": "2.1.3"
       },

package/examples/extensions/with-deps/package.json

@@ -1,7 +1,7 @@
 {
 	"name": "omk-extension-with-deps",
 	"private": true,
-	"version": "1.2.0",
+	"version": "1.2.3",
 	"type": "module",
 	"scripts": {
 		"clean": "echo 'nothing to clean'",

package/npm-shrinkwrap.json

@@ -1,17 +1,17 @@
 {
 	"name": "open-multi-agent-kit",
-	"version": "1.2.0",
+	"version": "1.2.3",
 	"lockfileVersion": 3,
 	"requires": true,
 	"packages": {
 		"": {
 			"name": "open-multi-agent-kit",
-			"version": "1.2.0",
+			"version": "1.2.3",
 			"license": "MIT",
 			"dependencies": {
-				"omk-agent-core": "^1.2.0",
-				"omk-ai": "^1.2.0",
-				"omk-tui": "^1.2.0",
+				"omk-agent-core": "^1.2.3",
+				"omk-ai": "^1.2.3",
+				"omk-tui": "^1.2.3",
 				"@silvia-odwyer/photon-node": "0.3.4",
 				"canonicalize": "4.0.0",
 				"chalk": "5.6.2",
@@ -27,8 +27,8 @@
 				"typebox": "1.3.11",
 				"undici": "8.10.0",
 				"yaml": "2.9.0",
-				"omk-adaptorch-wpl": "^1.2.0",
-				"omk-protocol": "^1.2.0"
+				"omk-adaptorch-wpl": "^1.2.3",
+				"omk-protocol": "^1.2.3"
 			},
 			"optionalDependencies": {
 				"@mariozechner/clipboard": "0.3.9"
@@ -1285,19 +1285,19 @@
 			}
 		},
 		"node_modules/omk-adaptorch-wpl": {
-			"version": "1.2.0",
-			"resolved": "https://registry.npmjs.org/omk-adaptorch-wpl/-/omk-adaptorch-wpl-1.2.0.tgz",
+			"version": "1.2.3",
+			"resolved": "https://registry.npmjs.org/omk-adaptorch-wpl/-/omk-adaptorch-wpl-1.2.3.tgz",
 			"license": "MIT",
 			"engines": {
 				"node": ">=22.19.0"
 			}
 		},
 		"node_modules/omk-agent-core": {
-			"version": "1.2.0",
-			"resolved": "https://registry.npmjs.org/omk-agent-core/-/omk-agent-core-1.2.0.tgz",
+			"version": "1.2.3",
+			"resolved": "https://registry.npmjs.org/omk-agent-core/-/omk-agent-core-1.2.3.tgz",
 			"license": "MIT",
 			"dependencies": {
-				"omk-ai": "^1.2.0",
+				"omk-ai": "^1.2.3",
 				"ignore": "7.0.6",
 				"typebox": "1.3.11",
 				"yaml": "2.9.0"
@@ -1307,8 +1307,8 @@
 			}
 		},
 		"node_modules/omk-ai": {
-			"version": "1.2.0",
-			"resolved": "https://registry.npmjs.org/omk-ai/-/omk-ai-1.2.0.tgz",
+			"version": "1.2.3",
+			"resolved": "https://registry.npmjs.org/omk-ai/-/omk-ai-1.2.3.tgz",
 			"license": "MIT",
 			"dependencies": {
 				"@anthropic-ai/sdk": "0.91.1",
@@ -1330,16 +1330,16 @@
 			}
 		},
 		"node_modules/omk-protocol": {
-			"version": "1.2.0",
-			"resolved": "https://registry.npmjs.org/omk-protocol/-/omk-protocol-1.2.0.tgz",
+			"version": "1.2.3",
+			"resolved": "https://registry.npmjs.org/omk-protocol/-/omk-protocol-1.2.3.tgz",
 			"license": "MIT",
 			"engines": {
 				"node": ">=22.19.0"
 			}
 		},
 		"node_modules/omk-tui": {
-			"version": "1.2.0",
-			"resolved": "https://registry.npmjs.org/omk-tui/-/omk-tui-1.2.0.tgz",
+			"version": "1.2.3",
+			"resolved": "https://registry.npmjs.org/omk-tui/-/omk-tui-1.2.3.tgz",
 			"license": "MIT",
 			"dependencies": {
 				"get-east-asian-width": "1.6.0",

package/package.json

@@ -1,6 +1,6 @@
 {
 	"name": "open-multi-agent-kit",
-	"version": "1.2.0",
+	"version": "1.2.3",
 	"description": "OMK (Open Multi-Agent Kit) coding agent CLI with read, bash, edit, write tools and session management",
 	"type": "module",
 	"omkConfig": {
@@ -43,9 +43,9 @@
 		"link-workspace-deps": "node ./scripts/link-workspace-deps.mjs"
 	},
 	"dependencies": {
-		"omk-agent-core": "^1.2.0",
-		"omk-ai": "^1.2.0",
-		"omk-tui": "^1.2.0",
+		"omk-agent-core": "^1.2.3",
+		"omk-ai": "^1.2.3",
+		"omk-tui": "^1.2.3",
 		"@silvia-odwyer/photon-node": "0.3.4",
 		"canonicalize": "4.0.0",
 		"chalk": "5.6.2",
@@ -61,8 +61,8 @@
 		"typebox": "1.3.11",
 		"undici": "8.10.0",
 		"yaml": "2.9.0",
-		"omk-adaptorch-wpl": "^1.2.0",
-		"omk-protocol": "^1.2.0"
+		"omk-adaptorch-wpl": "^1.2.3",
+		"omk-protocol": "^1.2.3"
 	},
 	"overrides": {
 		"rimraf": "6.1.2",
@@ -75,9 +75,7 @@
 	},
 	"devDependencies": {
 		"@types/cross-spawn": "6.0.6",
-		"@types/diff": "7.0.2",
 		"@types/hosted-git-info": "3.0.5",
-		"@types/ms": "2.1.0",
 		"@types/node": "24.12.4",
 		"@types/proper-lockfile": "4.1.4",
 		"fast-check": "4.9.0",