Tighten Scattermind rank feedback contract

This commit is contained in:
OpenClaw Bot
2026-05-26 22:17:47 +02:00
parent e532c6d910
commit 8b7477e323
3 changed files with 103 additions and 26 deletions
+40
View File
@@ -39,6 +39,46 @@ Tables:
- `milestones` — name, description, horizon, color, position, active
- `activity` — small append-only UX feed
## Scattermind → Ranker bridge
Ranker's continuation job is narrow:
`Snapshot / Concept Map → candidate feature/action set → Rank-ready build order`
`POST /api/rank-feedback` accepts a `prioritix-feature-set-v1`-style payload from Scattermind and returns ranked items plus `buildOrder.doFirst / validateNext / defer / park`. Keep this contract action-first; do not use it as a reason to add generic dashboard, auth, billing, or workspace layers before the bridge has proof.
Recommended payload shape:
```json
{
"schema": "prioritix-feature-set-v1",
"sourceName": "Scattermind",
"artifactId": "snapshot_or_concept_map_id",
"snapshotTitle": "Plain idea title",
"conceptMapId": "optional_concept_map_id",
"originalPrompt": "The user's starting prompt, trimmed for provenance",
"idea": "What Scattermind clarified",
"context": "Important constraints: solo builder, non-AI-native user, avoid dashboard swamp, etc.",
"mode": "mvp",
"featureSet": {
"features": [
{
"id": "build-order-preview",
"title": "Build order preview",
"description": "Show do first, validate next, defer, and park with reasons.",
"userValue": "A tired builder sees the next move without opening a dashboard.",
"evidenceNeeded": "Can 3 non-AI-native users understand the first recommended action?",
"proofSteps": ["Show a static result screen to 3 people"],
"dependencies": [],
"risk": "May become generic roadmap UI if the source context is lost.",
"recommendedLane": "validate-next",
"sourceSection": "concept-map.nextMoves"
}
]
}
}
```
## Commands
```bash