expxagents 0.24.0 → 0.24.1
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.
|
@@ -19,7 +19,14 @@ For each pipeline step:
|
|
|
19
19
|
|
|
20
20
|
#### Agent Steps (type: agent or default)
|
|
21
21
|
|
|
22
|
-
1. **
|
|
22
|
+
1. **Update state.json BEFORE executing:**
|
|
23
|
+
- Set `status` → `running`
|
|
24
|
+
- Set `step.current` → current step number (1-based)
|
|
25
|
+
- Set `step.label` → step label from squad.yaml
|
|
26
|
+
- Set agent `status` → `working`
|
|
27
|
+
- **CRITICAL: Set agent `stepIndex`** → current step number (1-based). This triggers the dashboard kanban to move the previous task to "done" and the current task to "executing". Without this update, kanban tasks won't transition.
|
|
28
|
+
- Set agent `stepLabel` → step label
|
|
29
|
+
- Write state.json immediately (before doing any work)
|
|
23
30
|
2. **Build context:**
|
|
24
31
|
- Agent prompt from `agents/<agent-id>.md`
|
|
25
32
|
- Previous agent output (if `deliverFrom` is set)
|
|
@@ -30,10 +37,10 @@ For each pipeline step:
|
|
|
30
37
|
- `inline` execution: run in current context
|
|
31
38
|
- `subagent` execution: spawn background process
|
|
32
39
|
4. **Save output** to `output/step-XX.md`
|
|
33
|
-
5. **Update state.json:**
|
|
40
|
+
5. **Update state.json AFTER executing:**
|
|
34
41
|
- Agent status → `delivering` (if handoff) or `done`
|
|
35
42
|
- Handoff message
|
|
36
|
-
-
|
|
43
|
+
- Write state.json immediately
|
|
37
44
|
6. **Handoff delay** (2 seconds for dashboard animation)
|
|
38
45
|
|
|
39
46
|
#### Checkpoint Steps (type: checkpoint)
|
|
@@ -50,28 +57,50 @@ For each pipeline step:
|
|
|
50
57
|
|
|
51
58
|
### State Management (state.json)
|
|
52
59
|
|
|
60
|
+
**IMPORTANT:** The dashboard file watcher monitors state.json for changes. When agent `stepIndex` changes, it automatically updates the kanban board (previous step → done, current step → executing). You MUST update `stepIndex` on each agent before starting their step.
|
|
61
|
+
|
|
53
62
|
```json
|
|
54
63
|
{
|
|
55
64
|
"squad": "<squad-code>",
|
|
56
65
|
"status": "running",
|
|
57
|
-
"step": { "current":
|
|
66
|
+
"step": { "current": 2, "total": 5, "label": "step-02-writing" },
|
|
58
67
|
"agents": [
|
|
59
68
|
{
|
|
60
69
|
"id": "researcher",
|
|
61
70
|
"name": "Angela Researcher",
|
|
62
71
|
"icon": "magnifying-glass",
|
|
63
|
-
"status": "
|
|
72
|
+
"status": "done",
|
|
73
|
+
"stepIndex": 1,
|
|
74
|
+
"stepLabel": "step-01-research",
|
|
64
75
|
"desk": { "col": 1, "row": 1 },
|
|
76
|
+
"deliverTo": "writer",
|
|
77
|
+
"message": "Research complete"
|
|
78
|
+
},
|
|
79
|
+
{
|
|
80
|
+
"id": "writer",
|
|
81
|
+
"name": "Bruno Writer",
|
|
82
|
+
"icon": "pen",
|
|
83
|
+
"status": "working",
|
|
84
|
+
"stepIndex": 2,
|
|
85
|
+
"stepLabel": "step-02-writing",
|
|
86
|
+
"desk": { "col": 2, "row": 1 },
|
|
65
87
|
"deliverTo": null,
|
|
66
88
|
"message": ""
|
|
67
89
|
}
|
|
68
90
|
],
|
|
69
91
|
"handoff": null,
|
|
70
92
|
"startedAt": "2026-03-13T00:00:00Z",
|
|
71
|
-
"updatedAt": "2026-03-13T00:00:
|
|
93
|
+
"updatedAt": "2026-03-13T00:00:05Z"
|
|
72
94
|
}
|
|
73
95
|
```
|
|
74
96
|
|
|
97
|
+
**Key fields for kanban integration:**
|
|
98
|
+
- `step.current` — global step number (1-based)
|
|
99
|
+
- `agent.stepIndex` — the step number this agent is currently on (1-based). Changes to this field trigger kanban task transitions
|
|
100
|
+
- `agent.stepLabel` — label of the current step
|
|
101
|
+
- `agent.status` — `working` | `delivering` | `done` | `idle` | `checkpoint`
|
|
102
|
+
```
|
|
103
|
+
|
|
75
104
|
### Output Versioning
|
|
76
105
|
|
|
77
106
|
- First run: `output/v1/`
|