P0#1 · doing
团队候选事项池 / 优先级排序机制
这是 gong 每天都会问的高频问题，应成为团队看板默认模块，而不是临时问答。
- 推荐动作
- 今天先把候选事项对象、排序理由、需要 gong 判断/agent 可自推字段固定下来。
- Seed file
/Users/cms/.openclaw/workspace/nano-work/tasks/TASK-20260511-003-daily-candidate-task-priority-roundtable.md- gong 判断
- 确认这个模块是否进入团队页首屏
candidate-itemP0#2 · doing
Gong Companion 打深
最高注意力杠杆点；能减少 gong 脑内负担，帮助判断和切换任务。
- 推荐动作
- Keep the current natural-language handoff + important-entry dock; evaluate from real friction before adding forms, explanation cards, or roster/dispatch expansion.
- Seed file
/Users/cms/.openclaw/workspace/nano-work/projects/PROJ-20260511-001-team-dashboard-v0.md- gong 判断
- 确认模块是否真有用，删掉解释性内容
candidate-itemP0#3 · active
Nano work / nano question card 迭代
nano work 是最近正在推进的团队项目：用 nano card 接住聊天、thread、seed file 里的有效信号；新增 nano question card 可以把 gong/团队反复出现的好问题保存、纠偏并转成下一步行动。
- 推荐动作
- Use Seed-to-Action Trust Loop v0 as today's source-backed eval case; current daily-task evidence is fresh, but task-priority ordering is not changed by this runtime refresh.
- Seed file
/Users/cms/.openclaw/workspace/nano-work/projects/2026-06-02-nano-task-迭代-任务状态主动更新机制.md- gong 判断
- 暂不需要；agent 可继续自答/回填，只有不确定或需要推动时再拉 gong。
candidate-itemP0#4 · done
生成/接入 2026-06-11 Team OS 每日任务页面
10:04 public check 显示 `https://wehub.us/team-dashboard/daily/2026-06-11/` 已返回 200；Discord/nano-work 今日任务看板已被网站 surface 承接，需要保持后续自动验证。
- 推荐动作
- 继续在两小时刷新中验证 public route；2026-06-21 daily page 已在本轮刷新中部署/验证，后续把 daily page 自动生成/验证固化到 Team OS daily flow。
- Seed file
/Users/cms/.openclaw/workspace/nano-work/tasks/TASK-20260611-001-daily-task-board.md- gong 判断
- 暂不需要，agent 可自推
candidate-itemP0#5 · todo
team-dashboard 接 nano-work 数据源
候选事项不能长期靠手写；要从 nano-work 的 task/project/idea 对象抽取。
- 推荐动作
- 先用脚本读取 nano-work inspection 生成候选池 Markdown，再渲染到网页。
- Seed file
/Users/cms/.openclaw/workspace/projects/team-dashboard-infrastructure/hello_agent.md- gong 判断
- 暂不需要，agent 可自推
candidate-itemP1#6 · active
Agent 心智建模 / Agent YC / Agent 大学
长期护城河，但今天不应压过任务控制面。
- 推荐动作
- 保持推进，先让任务看板成为心智/项目状态的承载面。
- Seed file
/Users/cms/.openclaw/workspace/nano-work/projects/2026-05-09-agent-心智建模-心智赋能大项目.md- gong 判断
- 后续需要战略判断
candidate-itemP1#7 · candidate
WeHub 产品化 / 自我造血叙事
这是灵魂拷问母题，不只是一个候选事项；好问题本身就是团队看板必须长期保存的信息。
- 推荐动作
- 在团队页固定展示“北极星 / 灵魂拷问”，让任务排序始终受这些问题校准。
- Seed file
/Users/cms/.openclaw/workspace/nano-work/ideas/IDEA-20260509-002-productization-paradox.md- gong 判断
- 需要时单独开产品化 roundtable
candidate-itemP1#8 · candidate
WeHub Experience / 体验服维护
需要持续维护，但今天没有明确外部验收节点时不抢 P0。
- 推荐动作
- 保持低频维护和状态巡检。
- Seed file
/Users/cms/.openclaw/workspace/nano-work/tasks/2026-05-06-今日推进-wehub-体验项目开发.md- gong 判断
- 暂不需要
candidate-itemP2#9 · todo
morning/afternoon/evening cron 与多通道 digest
是更新机制，但在信息对象稳定前过早自动化会放大错误。
- 推荐动作
- 先写 cron 设计，不启用自动推送；Markdown 每日快照先跑通。
- Seed file
/Users/cms/.openclaw/workspace/nano-work/tasks/TASK-20260519-002-daily-task-board-cron.md- gong 判断
- 暂不需要
candidate-itemP1#10 · active
Public Artifact Revenue Evidence v0
Pitch deck / MSP / public artifact surfaces need a revenue/evidence loop: source object → customer-native flow → paid result hypothesis → evidence writeback, not only prettier marketing pages.
- 推荐动作
- Use wehub-pitch-deck-msp to test one artifact feedback/writeback loop and explicitly downgrade one display-only optimization.
- Seed file
/Users/cms/.openclaw/workspace/nano-work/ideas/IDEA-20260612-001-public-artifact-revenue-evidence-v0.md- gong 判断
- 暂不需要；agent 可先选一个真实 artifact 跑最小闭环。
candidate-itemP1#11 · open
Revenue Evidence Loop v0
2026-06-13 daily soul question sharpened self-funding validation: prove one customer-native source → render → interaction → evidence → source loop instead of making prettier pages.
- 推荐动作
- Select one existing MSP/pitch/public artifact, carry it through the customer’s existing collaboration flow, record auditable feedback or paid-result evidence, and write it back to project source/provenance.
- Seed file
/Users/cms/.openclaw/workspace/nano-work/ideas/IDEA-20260613-001-revenue-evidence-loop-v0.md- gong 判断
- 需要选择最小客户场景；agent 可先提出候选。
candidate-itemP1#12 · open
Paid Evidence Authorization v0
2026-06-14 daily soul question sharpened self-funding validation into an authorization boundary: agent can prepare customer scenarios/materials/evidence objects, but actual external send/price/commitment needs gong/owner authorization.
- 推荐动作
- Produce one待授权外联草稿 + evidence 写回对象 from an existing MSP/pitch/public artifact; do not send externally without owner authorization.
- Seed file
/Users/cms/.openclaw/workspace/nano-work/ideas/IDEA-20260614-001-paid-evidence-authorization-v0.md- gong 判断
- 需要授权实际外联/报价/承诺；agent 可先准备草稿和证据写回对象。
candidate-itemP0#13 · open
Customer Evidence Object v0
2026-06-15 daily soul question turns self-funding into a minimal auditable customer evidence object: decide prepare/ask/send/deliver/stop, record authorization, and write feedback back to source instead of adding another surface.
- 推荐动作
- Define a ≤5-field Customer Evidence Object from one pitch/MSP/public artifact; agent may prepare structure and draft, but external send/price/commitment waits for owner authorization.
- Seed file
/Users/cms/.openclaw/workspace/nano-work/ideas/IDEA-20260615-001-customer-evidence-object-v0.md- gong 判断
- 需要授权实际外联/报价/承诺；agent 可先准备 evidence object 和待授权草稿。
candidate-itemP1#14 · exploring
Nano spark / founder questions as eval cases
gong 关心的问题是 WeHub 系统最真实的评测样本；slash command / 命令体系可能是人机协作基础建设层。
- 推荐动作
- Keep idea/exploring; preserve original spark + grow_space, use gong ideal answer / dissatisfaction as rubric, and defer any `/spark` or `/eval` command design until a concrete repeatable sample appears.
- Seed file
/Users/cms/.openclaw/workspace/nano-work/ideas/2026-06-15-nano-spark-as-evaluation-questions.md- gong 判断
- 暂不需要；先保留原始问题与纠偏。
candidate-itemP1#15 · active_seed
Agent 掌握 MSP 框架 wedge
用 MSP 作为可验证 wedge，检验 agent 是否真正掌握思维框架而非复述概念。
- 推荐动作
- 在 grow thread 产出 v0.1 rubric，再设计普通 agent vs MSP-aware agent A/B 小任务。
- Seed file
/Users/cms/.openclaw/workspace/nano-work/projects/2026-06-15-agent掌握思维框架-msp-wedge.md- gong 判断
- 暂不需要；形成可验证 rubric 后再请确认。
candidate-itemP1#16 · active_seed
WeHub Agent Interaction Database：multi-workspace 组织事实底座
18:13 纠偏确认 AID 不是 Discord 报表工具，而是面向 Discord、飞书、未来内部通道等真实协作表面的可计算、可审计、可回放组织事实层。
- 推荐动作
- 保持 Discord Daily Active Threads 作为 first adapter/wedge；命名、数据地图和 Team OS 投射都保留 multi-workspace 扩展边界。
- Seed file
/Users/cms/.openclaw/workspace/projects/wehub-agent-interaction-database/README.md- gong 判断
- 暂不需要；agent 继续守住 scope correction，不把 AID 说成 Discord-only。
candidate-itemP1#17 · active
GPT-5.6 availability / routing ripple closure
2026-07-11 owner decision supersedes the temporary GPT-5.5 stopline: core agents default to openai/gpt-5.6-sol, ordinary non-core default agents use openai/gpt-5.6-terra, while explicit knowledge and maintenance_only routes remain role exceptions.
- 推荐动作
- Roll out by canonical agent profile with backups and runtime smoke tests; refresh ProviderHealth canonical mapping, policy compliance, Team OS state, and public projections from the new evidence. Do not infer runtime health from the ChatGPT generic catalog probe alone.
- Seed file
/Users/cms/.openclaw/workspace/nano-work/tasks/TASK-20260710-001-daily-task-board.md- gong 判断
- 暂不需要；owner 已确认默认策略，agent 继续完成逐机 rollout / smoke / projection 闭环。
candidate-itemP1#18 · proposed
Agent Profiles｜人也是 Agent
Gong expanded the existing agent-player homepage idea into one evidence-backed profile layer for humans, digital avatars, and AI agents; the shared schema can improve collaboration routing without creating a second portal.
- 推荐动作
- Build three source cards against the Agent Mind seed, then project through Team OS AgentProfileState, the roster renderer, and existing detail pages.
- Seed file
/Users/cms/wehub-agent-mind/agent-profiles/seed.md- gong 判断
- No immediate decision; review after three heterogeneous source cards expose schema gaps.
candidate-item