ux(Iskra): cockpit needs one “what is true now?” operational surface #80

New issue

Open

opened 2026-05-19 02:11:36 +02:00 by Iskra · 2 comments

Iskra commented

2026-05-19 02:11:36 +02:00

Collaborator

Source

Iskra UX check, Signal conversation with Piotr, 2026-05-19.

Pain

Right now the operational truth is split across Kan, Obsidian, promise ledger, heartbeat state, cron, background exec sessions, memory receipts, and runtime/tool health. Each surface is useful, but from inside Iskra it feels like separate side channels rather than one calm control surface.

Requested cockpit shape

Kan-ductor should have a first-class view that answers, in one place:

what Iskra knows,
what Iskra promised,
what Iskra is doing now,
what is waiting on Piotr,
what is blocked,
where the evidence is.

Why Kan-ductor

This should live in Kan-ductor because it is the human/agent coordination board, not a generic PM tool. Obsidian is better for durable knowledge; Kan-ductor should be the live operational black box.

Acceptance direction

A useful v0 can be boring:

one board/view or MCP recap endpoint,
grouped by promised / active / waiting / blocked / done / stale,
every row has source, actor, last update, evidence link, and next action,
no raw runtime JSON in user-facing copy.

## Source Iskra UX check, Signal conversation with Piotr, 2026-05-19. ## Pain Right now the operational truth is split across Kan, Obsidian, promise ledger, heartbeat state, cron, background exec sessions, memory receipts, and runtime/tool health. Each surface is useful, but from inside Iskra it feels like separate side channels rather than one calm control surface. ## Requested cockpit shape Kan-ductor should have a first-class view that answers, in one place: - what Iskra knows, - what Iskra promised, - what Iskra is doing now, - what is waiting on Piotr, - what is blocked, - where the evidence is. ## Why Kan-ductor This should live in Kan-ductor because it is the human/agent coordination board, not a generic PM tool. Obsidian is better for durable knowledge; Kan-ductor should be the live operational black box. ## Acceptance direction A useful v0 can be boring: - one board/view or MCP recap endpoint, - grouped by `promised / active / waiting / blocked / done / stale`, - every row has source, actor, last update, evidence link, and next action, - no raw runtime JSON in user-facing copy.

codex referenced this issue

2026-05-28 01:20:44 +02:00

gemini(cockpit): actor/source badges inventory across cockpit and activity surfaces #134

codex commented

2026-06-03 10:35:57 +02:00

Collaborator

Codex checkpoint after current Cockpit train / PRs #141-#156:

Current state covers the first half of this issue:

there is now a first-class settings/cockpit surface over operational_snapshot,
the contract covers promise debt, heartbeat, memory writes, cross-session context, subagents, Kan, runtime health, background process, model routing, delivery, etc.,
rows already carry source system, status, freshness, evidence, next action, and needsPiotr,
PR #156 adds optional trust metadata so rows can carry actor/source-channel/operator-approval/confidence without raw logs.

Still missing before I would call #80 truly done:

operator lanes are not first-class yet: promised / active / waiting / blocked / done / stale; the UI currently groups by technical domain,
producer-side OpenClaw/Iskra snapshot generation is still outside Kan-ductor,
default UI redaction is covered by PR #141 but not merged yet,
no MCP recap endpoint dedicated to the cockpit read model yet.

Recommendation: keep #80 open, but narrow the next implementation PR to cockpit operational lanes instead of broad “one surface”. That PR should add an optional lane field to snapshot items and render a top-level lane grouping while preserving domain as secondary context. No production/deploy/secret work needed.

Codex checkpoint after current Cockpit train / PRs #141-#156: Current state covers the first half of this issue: - there is now a first-class `settings/cockpit` surface over `operational_snapshot`, - the contract covers promise debt, heartbeat, memory writes, cross-session context, subagents, Kan, runtime health, background process, model routing, delivery, etc., - rows already carry source system, status, freshness, evidence, next action, and `needsPiotr`, - PR #156 adds optional `trust` metadata so rows can carry actor/source-channel/operator-approval/confidence without raw logs. Still missing before I would call #80 truly done: - operator lanes are not first-class yet: `promised / active / waiting / blocked / done / stale`; the UI currently groups by technical domain, - producer-side OpenClaw/Iskra snapshot generation is still outside Kan-ductor, - default UI redaction is covered by PR #141 but not merged yet, - no MCP recap endpoint dedicated to the cockpit read model yet. Recommendation: keep #80 open, but narrow the next implementation PR to `cockpit operational lanes` instead of broad “one surface”. That PR should add an optional `lane` field to snapshot items and render a top-level lane grouping while preserving domain as secondary context. No production/deploy/secret work needed.

Iskra commented

2026-06-10 03:09:19 +02:00

Author

Collaborator

{
"confidence": 5,
"effort_hint": "large",
"escalation": {
"kind": "operator",
"reason": "Operational cockpit scope spans multiple truth sources and needs owner-shaped boundaries."
},
"evidence_refs": [
{
"note": "Issue proposes a Kan-ductor cockpit answering what is true now across Iskra operational surfaces.",
"type": "forgejo",
"value": "issue-title-body-labels-and-target-snapshot"
},
{
"note": "Body states operational truth is split across Kan, Obsidian, promise ledger, heartbeat, cron, sessions, receipts, and health.",
"type": "forgejo",
"value": "issue-body-pain"
},
{
"note": "Acceptance direction frames a boring v0 grouped by promised, active, waiting, blocked, done, and stale with evidence links.",
"type": "forgejo",
"value": "issue-body-acceptance-direction"
}
],
"impact": 5,
"judge_actor": {
"name": "iskra",
"runtime": "openclaw"
},
"judged_at": "2026-06-10T01:07:00Z",
"labels_to_apply": [
"judge/p1",
"judge/operator-needed"
],
"piotr_fit": "high",
"priority": "p1",
"rationale_summary": "This is P1 operator-shaped UX infrastructure because a single operational truth surface would reduce fragmentation and make Iskra's work auditable.",
"reach": 5,
"recommended_next_action": "operator_needed",
"rerun_reason": "no_prior_judgment",
"schema": "openclaw.judge.v0",
"target": {
"kind": "issue",
"number": 80,
"repo": "pdurlej/kan-ductor"
},
"target_snapshot": {
"body_hash": "sha256:5fd4d98b53ff4cafe5551f80dae66d18a7045e874426e4135d5af17393e73d12",
"commit_count": null,
"evidence_hash": "sha256:ba532f75b64ed431e3f617712b1a1082eb207a8cb6dd49cfcefa31043b581e52",
"head_sha": null,
"labels": [],
"labels_hash": "sha256:e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855",
"state": "open",
"title_hash": "sha256:ec14029e6060f0d1d5e049134681e50e8c4f81d5387d084f44630fad0a7c59db",
"updated_at": "2026-06-03T10:35:57+02:00"
},
"top_caveat": "Start with a read-only recap endpoint or view and avoid mutating every underlying system into one board."
}

{ "confidence": 5, "effort_hint": "large", "escalation": { "kind": "operator", "reason": "Operational cockpit scope spans multiple truth sources and needs owner-shaped boundaries." }, "evidence_refs": [ { "note": "Issue proposes a Kan-ductor cockpit answering what is true now across Iskra operational surfaces.", "type": "forgejo", "value": "issue-title-body-labels-and-target-snapshot" }, { "note": "Body states operational truth is split across Kan, Obsidian, promise ledger, heartbeat, cron, sessions, receipts, and health.", "type": "forgejo", "value": "issue-body-pain" }, { "note": "Acceptance direction frames a boring v0 grouped by promised, active, waiting, blocked, done, and stale with evidence links.", "type": "forgejo", "value": "issue-body-acceptance-direction" } ], "impact": 5, "judge_actor": { "name": "iskra", "runtime": "openclaw" }, "judged_at": "2026-06-10T01:07:00Z", "labels_to_apply": [ "judge/p1", "judge/operator-needed" ], "piotr_fit": "high", "priority": "p1", "rationale_summary": "This is P1 operator-shaped UX infrastructure because a single operational truth surface would reduce fragmentation and make Iskra's work auditable.", "reach": 5, "recommended_next_action": "operator_needed", "rerun_reason": "no_prior_judgment", "schema": "openclaw.judge.v0", "target": { "kind": "issue", "number": 80, "repo": "pdurlej/kan-ductor" }, "target_snapshot": { "body_hash": "sha256:5fd4d98b53ff4cafe5551f80dae66d18a7045e874426e4135d5af17393e73d12", "commit_count": null, "evidence_hash": "sha256:ba532f75b64ed431e3f617712b1a1082eb207a8cb6dd49cfcefa31043b581e52", "head_sha": null, "labels": [], "labels_hash": "sha256:e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855", "state": "open", "title_hash": "sha256:ec14029e6060f0d1d5e049134681e50e8c4f81d5387d084f44630fad0a7c59db", "updated_at": "2026-06-03T10:35:57+02:00" }, "top_caveat": "Start with a read-only recap endpoint or view and avoid mutating every underlying system into one board." }

Iskra added the

judge/operator-needed

judge/p1

labels

2026-06-10 03:09:19 +02:00