

























@@ -9,7 +9,8 @@ title: "Workboard plugin"
991010The Workboard plugin adds an optional Kanban-style board to the
1111[Control UI](/web/control-ui). Use it to collect agent-sized work cards, assign
12-them to agents, and jump from a card into the linked dashboard session.
12+them to agents, and track the linked background task, run, and dashboard
13+session from one card.
13141415Workboard is intentionally small. It tracks local operating work for an
1516OpenClaw Gateway; it is not a replacement for GitHub Issues, Linear, Jira, or
@@ -47,8 +48,8 @@ Each card stores:
4748- priority: `low`, `normal`, `high`, or `urgent`
4849- labels
4950- optional agent id
50-- optional linked session, run, task, or source URL
51-- optional execution metadata for a Codex or Claude session started from the card
51+- optional linked task, run, session, or source URL
52+- optional execution metadata for a Codex or Claude run started from the card
5253- compact metadata for attempts, comments, links, proof, artifacts, automation,
5354 attachments, worker logs, worker protocol state, claims, diagnostics,
5455 notifications, templates, archive state, and stale-session detection
@@ -65,26 +66,35 @@ proof snippets, related links, comments, archive markers, and stale-session
6566markers are intentionally local metadata; they do not replace session
6667transcripts or GitHub issue history.
676868-## Card executions
69+## Card executions and tasks
697070-Unlinked cards can start work from the card. Start uses the Gateway's configured
71+Unlinked cards can start work from the card. Autonomous starts use the
72+Gateway's task-tracked agent run path, then Workboard links the resulting task,
73+run id, and session key back onto the card. Start uses the Gateway's configured
7174default agent and model. Codex and Claude actions are optional explicit model
7275choices:
737674-- Run Codex or Run Claude creates a dashboard session, sends the card prompt,
75- and marks the card `running`.
77+- Run Codex or Run Claude starts a task-backed agent run, sends the card
78+prompt, and marks the card `running`.
7679- Open Codex or Open Claude creates a linked dashboard session without sending
7780 the card prompt or moving the card, so you can work manually while it stays
7881 attached to the board.
79828083Execution metadata stores the selected engine, mode, model ref, session key,
81-run id, and lifecycle status on the card. Codex executions use
82-`openai/gpt-5.5`; Claude executions use `anthropic/claude-sonnet-4-6`.
84+run id, task id when available, and lifecycle status on the card. Codex
85+executions use `openai/gpt-5.5`; Claude executions use
86+`anthropic/claude-sonnet-4-6`.
83878488Each linked execution also records an attempt summary on the same card record.
8589The attempt summary keeps the engine, mode, model, run id, timestamps, status,
8690and rolling failure count so repeated failures remain visible on the board.
879192+The dashboard refreshes task status from the Gateway task ledger and matches
93+tasks back to cards by task id, run id, or linked session key. If a task is
94+queued or running, the card lifecycle shows active task state. If the task
95+finishes, fails, times out, or is cancelled, the card lifecycle moves toward
96+review or blocked status using the same lifecycle sync as linked sessions.
97+8898## Agent coordination
899990100Workboard also exposes optional agent tools for board-aware workflows:
@@ -160,13 +170,15 @@ blocked cards that need attention, repeated failures, done cards without proof,
160170and running cards that only have a loose session link.
161171162172Dispatch is intentionally Gateway-local. It does not spawn arbitrary operating
163-system processes; normal OpenClaw subagent sessions still own execution. A
164-dispatch nudge promotes dependency-ready cards, records dispatch metadata on
173+system processes; normal OpenClaw subagent sessions still own execution. The
174+dispatch action promotes dependency-ready cards, records dispatch metadata on
165175ready cards, blocks expired claims or timed-out runs, marks board-configured
166176triage cards as orchestration candidates, then claims a small batch of ready
167-cards and starts worker runs through the Gateway subagent runtime. Workers get
168-bounded card context plus the claim token they need to heartbeat, complete, or
169-block the card through the Workboard tools.
177+cards and starts worker runs through the Gateway subagent runtime. Assigned
178+cards use `agent:<id>:subagent:workboard-*` worker session keys; unassigned
179+cards use unscoped `subagent:workboard-*` keys so the Gateway still resolves the
180+configured default agent. Workers get bounded card context plus the claim token
181+they need to heartbeat, complete, or block the card through the Workboard tools.
170182171183### Dispatch worker selection
172184此内容由惯性聚合(RSS阅读器)自动聚合整理,仅供阅读参考。 原文来自 — 版权归原作者所有。