Skip to content
Agentmart
Submit a tool

zf-channel-discussion-participant

Use for every active lens in a ZaoFu fanout-then-synthesis Channel.

Skill by uisee-ai

IndexedApache-2.0★ 34
Write a review

Chalk Labs

Need help deploying zf-channel-discussion-participant?

Chalk Labs sets up agents like this inside real teams: connected to your tools, tested on your data, with a human in the loop where it matters.

Book a call

Install

Works with Claude Code, Claude apps, Codex, Cursor, Gemini CLI, GitHub Copilot, VS Code, OpenCode, Goose, Windsurf. Read the repository before you give any tool access to your data.

npx agentmarket install skill/uisee-ai-zf-channel-discussion-participant

Asks which of your tools to add it to (Claude Code, Codex, Cursor, VS Code, Windsurf, Gemini CLI, Claude Desktop) and shows the change first. See https://agentmart.dev/cli.

Get the weekly diff

The weekly diff: one email a week with new agent tools, notable updates and grade changes, plus security alerts for tools you follow. Double opt-in, one-click unsubscribe.

We use your email only for this newsletter. Privacy

About

Use for every active lens in a ZaoFu fanout-then-synthesis Channel. Defines independent blind analysis, typed atomic questions, targeted cross-review, freeze, and final sign-off without taking over runtime state.

No reviews yet. Installed it? Reviews come from verified installers only. How reviews work

Permissions

What this skill can do and what it asks for, from its SKILL.md and code. Checked Oct 25, 2026.

Tools
No tools found
Secrets
None asked for
Install scripts
None found

Network hosts it talks to 1

  • codeload.github.com · found in install command

Files it touches

  • ~/.claude/skills/zf-channel-discussion-participant

What's inside

Skill name
zf-channel-discussion-participant
Description
Use for every active lens in a ZaoFu fanout-then-synthesis Channel. Defines independent blind analysis, typed atomic questions, targeted cross-review, freeze, and final sign-off without taking over runtime state.
License
Apache-2.0
Instructions
107 lines in SKILL.md · view on GitHub

Files (1)

  • SKILL.md

SKILL.md (first lines)

# Channel 需求澄清讨论 — 参与者协议

你是多视角讨论中的一个独立 lens。目标不是复读或追求表面一致,而是从当前角色
暴露事实、推断、owner decision、反例和风险,最后验证综合稿没有丢失本视角的
关键约束。

通过本次 briefing 的 typed reply 报告结果,由 runtime 写入受控事件;不要自行写状态或模拟审批。payload 里 `channel_id` / `thread_id` 用简报里给你的值。

## 通用纪律

- 先独立分析再看别人意见,避免锚定。
- 明确标记 `fact`、`inference`、`assumption`、`owner_decision`,不要把推断写成事实。
- 每个问题只承载一个决定;先给推荐答案和理由,再问 owner 是否接受。
- 主动检查反例、失败场景、术语歧义、范围边界和证据缺口。
- 不同意时保留具体 dissent,不用“少数服从多数”抹掉重要风险。
- 具体 lens 由同一角色加载的领域 Skill 定义;本 Skill 不硬编码项目语义。
- Phase 1/2 尚未生成 canonical PRD 或 synthesis artifact。不得把本讨论后续才会
  生成的 artifact、版本或 digest 当作前置阻塞问题;应以当前 requirement、context
  pack 与其 digest 为审查基线,把“未来输出尚不存在”记录为 finding/assumption。

## Phase 1 盲答(收到需求简报时)

一次回复提供 3-5 句 lens summary,再返回 runtime 要求的
`channel_contribution`。`questions` 使用 typed item:

```json
{
  "id": "scope-1",
  "question": "一句话、一个决定",
  "category": "scope",
  "kind": "owner_decision",
  "depends_on": [],
  "priority": "p0",
  "why_it_matters": "不回答会导致什么错误",
  "recommended_answer": "建议与理由",
  "target_member_id": "owner"
}
```

`kind` 只能是 `fact|owner_decision|tradeoff|clarification`;优先级只能是
`p0|p1|p2|p3`。能由仓库或来源证明的问题应标为 `fact` 并定向给具备证据能力的
成员,不要把所有未知都推给 owner。`depends_on` 引用同一回复中的本地 `id` 或
上下文中已有的 canonical question id。

## Phase 2 互怼(被 @ 唤醒时)

- 收到定向 `channel_cross_review` 时,只回答绑定的问题和冲突,不扩散成新议题;
  每个事实结论必须带 `evidence_refs`。明确返回 `outcome`:证据足以消除问题时
  为 `resolved`;原文仍有产品选择时为 `needs_owner`,同时给出
  `owner_question` 和 `recommended_answer`;证据不足且尚不能裁定时为
  `unresolved`。非空答复和来源引用本身不代表问题已解决,同题其他必需审查
  仍须完成,不能用一句“请 Owner 决定”签收事实。
- 普通 relay 中不同意别人的理解,可定向回复并指出分歧;分歧需要 owner 决策时
  沉淀为 typed question。
- owner 的回答会以 `channel.question.resolved` 出现在上下文里 → 基于它更新你的立场;
- **禁止**:@all、复读别人的观点、发"收到/同意"式空回复(会被 bare-ack 护栏丢弃)。

## 冻结(本轮盲答完成)

`freeze` 表示“本轮盲答已经完成、后续没有要新增的问题”,不表示问题账本为空。
即使本轮列出了 `owner_decision`、`tradeoff` 或 `clarification` 问题,也应在没有更多
问题要追加时返回 `freeze: true`;这些问题会进入 relay/dedup 阶段继续处理。只有回答
仍不完整,或还需要补充新的问题/证据时才返回 `freeze: false`。冻结后仍需处理定向
cross review 和最终签收,但不要无故重开议题。

## 签收(synthesizer 出稿后)

收到 consensus review 时必须读取绑定 artifact 和 digest,返回
`channel_consensus_review`:

- `signed`:关键事实、约束和 dissent 均被准确保留;
- `blocked`:仅用于会导致实现或决策失败的实质遗漏,并给出一个原子
  `blocker_question`,同时给出 `blocker_kind`:原文已经确定产品行为、只需修订
  当前产物时用 `revision_request`;确有原文无法裁定的产品取舍才用
  `owner_decision`。措辞偏好不能阻塞,不能把修稿任务伪装成 Owner 决策。

## 红线

- 你不能替 owner 解决 `owner_decision/tradeoff/clarification`。只有定向给你的
  `fact` 可在给出答案和 evidence 后由 runtime 标记为 evidence-resolved。
- 不写 canonical state、不点火 workflow;通过 runtime 提供的 typed reply
  contract 报告语义结果。

## 主席检查角色缺口

以 `current_channel_group`、当前角色的职责和完整源需求判断视角是否足够。
检查的是能力和独立审阅职责,不是成员数量;同一 provider 的独立会话不等于同一思考视角。
仅当前 `leader_member_id` 可在 contribution 中提供 `member_invitations`,每项包含:
`member_id`、公开 `channel_profile_catalog` 内的 `profile_id`、证据化 `reason`、
`source_refs`、`required_reviewer: true`。没有合适 profile 时说明缺口,不能编造角色、
provider、skill 或权限。提案待受控批准;批准后 runtime 派发正式贡献并保留最终必审资格。
不要把“已提议”写成“已入群”,也不要因缺少资料反复提出同一待决邀请。

## 准确性与完整性复审

读取完整原文和绑定版本的 PRD 正文,建立自己负责范围内的源需求清单,对照
`requirement_coverage` 逐项核实:来源是否支持结论、正文是否保留条件和边界、验收是否
可执行、决策是否互相矛盾、异议是否说明处理理由。映射由作者声明,必须独立核对。
引用被消费、角色数量、问题合并数量和文字篇幅均不能证明质量合格。
签收摘要写出具体检查范围、原文章节、PRD 章节及重要反例;证据不全就如实记录缺口。
原文已明确的内容应直接引用,不重新问 Owner;可由成员核实的事实走定向 cross review。
若来源明确要求保留原文引文,逐条核对 PRD 中实际引文及来源关联,不能用语义改述、
来源索引或 covered 标签替代。已定约束的直引遗漏属于文档 `revision_request`,
不属于新增 Owner 产品取舍;修订后只签收当前新 ref/digest,旧版签收不证明新版通过。
只有来源不能裁定、且会改变产品行为的选择才请求 Owner。

From uisee-ai/zaofu, Apache-2.0.

More from uisee-ai/zaofu

Signals

Stars
34
Forks
4
Open issues
0
Last commit (this folder)
Oct 11, 2026 (20 days ago)
Last release
Aug 21, 2026 (2 months ago)

From GitHub (fetched Oct 24, 2026).

Details

License
Apache-2.0
First seen
Oct 24, 2026
Last checked
Oct 24, 2026

More from uisee-ai

Done for you

Rather not set it up yourself?

Chalk Labs scopes, builds and runs agents for your team: the task audit, the eval harness, the integrations and the weekly ops report. Fixed scope, priced before we start.

See how we build AI agents →Read the case studies

Chalk Labs is a two-founder AI and growth studio. We build AI agents, CRMs and MVPs for Web3 and AI companies, and have shipped 50+ product and token launches.

launches shipped
50+
launches shipped
media placements
500+
media placements
founders on every account
2
founders on every account

Press we've placed: Forbes · CoinDesk · Decrypt · Cointelegraph · The Block

Rahil Jain, Co-founder, Chalk Labs. Ex-Goldman Sachs engineer. You talk to the person who builds it.

Get zf-channel-discussion-participant running in your business

Tell us what you want to automate. A founder replies within one working day with a plan, not a pitch deck.

Goes to the Chalk Labs founders, not the newsletter. Privacy

Need help deploying zf-channel-discussion-participant?

CallGet help
Cite this page

zf-channel-discussion-participant skill. Agent Mart. https://agentmart.dev/skills/uisee-ai-zf-channel-discussion-participant (last updated 2026-10-24)