@enderfga/claw-orchestrator 7.5.4 → 7.5.6
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- package/README.md +3 -3
- package/dist/src/embedded-server.js +1 -1
- package/dist/src/embedded-server.js.map +1 -1
- package/dist/src/index.js +2 -2
- package/dist/src/index.js.map +1 -1
- package/dist/src/kernel/engine.d.ts +39 -1
- package/dist/src/kernel/engine.js +120 -10
- package/dist/src/kernel/engine.js.map +1 -1
- package/dist/src/kernel/types.d.ts +7 -0
- package/dist/src/kernel/types.js.map +1 -1
- package/dist/src/models.js +3 -3
- package/dist/src/models.js.map +1 -1
- package/dist/src/persistent-agy-session.d.ts +7 -0
- package/dist/src/persistent-agy-session.js +45 -8
- package/dist/src/persistent-agy-session.js.map +1 -1
- package/dist/src/session-manager.d.ts +7 -2
- package/dist/src/session-manager.js +7 -2
- package/dist/src/session-manager.js.map +1 -1
- package/package.json +1 -1
- package/skills/references/claude-cli-tracking.md +2 -1
- package/skills/references/multi-engine.md +10 -6
- package/skills/references/tools.md +13 -10
- package/skills/references/ultra.md +6 -6
- package/skills/references/workflow.md +12 -3
|
@@ -39,7 +39,7 @@ Start a persistent coding session with full CLI flag support.
|
|
|
39
39
|
| `betas` | string \| string[] | Custom beta headers |
|
|
40
40
|
| `enableAgentTeams` | boolean | Enable experimental agent teams |
|
|
41
41
|
| `enableAutoMode` | boolean | Enable auto permission mode |
|
|
42
|
-
| `customEngine` | object | Custom engine config (required when `engine='custom'`). See [Multi-Engine: Custom Engine](./multi-engine.md#custom-engine-engine-custom).
|
|
42
|
+
| `customEngine` | object | Custom engine config (required when `engine='custom'`). See [Multi-Engine: Custom Engine](./multi-engine.md#custom-engine-engine-custom). |
|
|
43
43
|
| `crossSessionInbound` | string | `accept` / `hold` / `refuse` — policy for peer messages from other Claude Code sessions on this machine (Claude engine). Delivered as a settings key; there is no CLI flag. Without it the CLI holds messages whose two sides run different permission modes, which is the usual orchestrated-session-to-human-terminal case |
|
|
44
44
|
| `includeHookEvents` | boolean | Stream hook lifecycle events (PreToolUse/PostToolUse) as `system` events |
|
|
45
45
|
| `forwardSubagentText` | boolean | Forward subagent text and thinking into the output stream (Claude engine, CLI 2.1.211+). Without it the parent stream stays quiet while a subagent works |
|
|
@@ -540,13 +540,13 @@ Get status and plan text when completed.
|
|
|
540
540
|
|
|
541
541
|
Start 1–20 reviewer agents that review the code in parallel, each from a different angle. Runs in background.
|
|
542
542
|
|
|
543
|
-
| Parameter | Type | Required | Description
|
|
544
|
-
| -------------------- | -------- | -------- |
|
|
545
|
-
| `cwd` | string | yes | Project directory
|
|
546
|
-
| `agentCount` | number | | Agents (1-20, default 5)
|
|
547
|
-
| `maxDurationMinutes` | number | | Duration (5-25 min, default 10)
|
|
548
|
-
| `model` | string | | Model for reviewers
|
|
549
|
-
| `focus` | string | | Review focus area
|
|
543
|
+
| Parameter | Type | Required | Description |
|
|
544
|
+
| -------------------- | -------- | -------- | -------------------------------------------------------------------------------------------------------------------------------- |
|
|
545
|
+
| `cwd` | string | yes | Project directory |
|
|
546
|
+
| `agentCount` | number | | Agents (1-20, default 5) |
|
|
547
|
+
| `maxDurationMinutes` | number | | Duration (5-25 min, default 10) |
|
|
548
|
+
| `model` | string | | Model for reviewers |
|
|
549
|
+
| `focus` | string | | Review focus area |
|
|
550
550
|
| `engines` | string[] | | Engines to spread reviewers across (default `["claude"]`). Every reviewer runs read-only, so `grok` and `custom` are not allowed |
|
|
551
551
|
|
|
552
552
|
### `ultrareview_status`
|
|
@@ -830,8 +830,11 @@ same as a failure).
|
|
|
830
830
|
trust.
|
|
831
831
|
- `workflow_cancel({ runId })`.
|
|
832
832
|
- `workflow_steer({ runId, text })` — the text is **prepended** to the next agent
|
|
833
|
-
node's prompt.
|
|
834
|
-
- `workflow_approve({ runId, approved })` — answers a `human_gate` node.
|
|
833
|
+
node's prompt. Steers not yet consumed are delivered after a restart too.
|
|
834
|
+
- `workflow_approve({ runId, approved })` — answers a `human_gate` node. A run
|
|
835
|
+
parked before the server restarted is resumed with the answer attached, so it
|
|
836
|
+
needs no separate `workflow_resume`. Returns `{ answered: false }` when the run
|
|
837
|
+
is not parked at a gate.
|
|
835
838
|
|
|
836
839
|
### `verify_run`
|
|
837
840
|
|
|
@@ -116,12 +116,12 @@ if (status?.status === 'completed') {
|
|
|
116
116
|
|
|
117
117
|
### Configuration
|
|
118
118
|
|
|
119
|
-
| Parameter | Default | Range | Description
|
|
120
|
-
| -------------------- | ------------------------- | ----- |
|
|
121
|
-
| `agentCount` | 5 | 1-20 | Number of reviewer agents
|
|
122
|
-
| `maxDurationMinutes` | 10 | 5-25 | Per-agent timeout
|
|
123
|
-
| `model` | session default | — | Model for all reviewers
|
|
124
|
-
| `focus` | bugs + security + quality | — | Review focus description
|
|
119
|
+
| Parameter | Default | Range | Description |
|
|
120
|
+
| -------------------- | ------------------------- | ----- | ----------------------------------------------------------------------------------------------------- |
|
|
121
|
+
| `agentCount` | 5 | 1-20 | Number of reviewer agents |
|
|
122
|
+
| `maxDurationMinutes` | 10 | 5-25 | Per-agent timeout |
|
|
123
|
+
| `model` | session default | — | Model for all reviewers |
|
|
124
|
+
| `focus` | bugs + security + quality | — | Review focus description |
|
|
125
125
|
| `engines` | `['claude']` | — | Engines assigned to reviewers round-robin. Not `grok`, which refuses a read-only session, or `custom` |
|
|
126
126
|
|
|
127
127
|
Reviewers run once each (up to 20 turns); there is no cross-review round. Results are stored as a durable run and remain queryable after a restart.
|
|
@@ -35,9 +35,18 @@ total: if `run.json` is missing or half-written, state is rebuilt by replaying
|
|
|
35
35
|
`events.jsonl` against `spec.json`. The atomic rewrite makes that path rare; the
|
|
36
36
|
replay makes it survivable anyway.
|
|
37
37
|
|
|
38
|
-
A kernel resumes at a **node boundary**, never mid-node
|
|
39
|
-
|
|
40
|
-
|
|
38
|
+
A kernel resumes at a **node boundary**, never mid-node, and at the node the run
|
|
39
|
+
was on — not at the first pending node, since a node on a branch the run never
|
|
40
|
+
took stays pending. Nodes already marked succeeded are not re-run; the node that
|
|
41
|
+
was in flight when the process died is retried from the start, because a
|
|
42
|
+
half-finished node left no result to trust. A node that had finished just before
|
|
43
|
+
the process died continues at its successor; a router is evaluated again, since
|
|
44
|
+
its choice is not recorded. A run parked at a `human_gate` parks there again.
|
|
45
|
+
|
|
46
|
+
Steers survive a restart. They are logged when they arrive and recorded as
|
|
47
|
+
consumed when the node that took them finishes, so on resume every steer no
|
|
48
|
+
finished node consumed is queued again — including one a node was holding when
|
|
49
|
+
it died, which goes to that node's retry.
|
|
41
50
|
|
|
42
51
|
**This makes node execution at-least-once, not exactly-once.** There is no
|
|
43
52
|
idempotency key and no side-effect commit marker, so a node that wrote files and
|