fyaic · veil-chow-fyaic · Sep 1, 2026 · Sep 1, 2026
@@ -0,0 +1,63 @@
+name: 15-minute operator report
+description: Report an independent first-run attempt of the ThreadMesh proof.
+title: "[Operator]: "
+labels: [external-review, usability, triage]
+body:
+  - type: input
+    id: environment
+    attributes:
+      label: Harness, OS, and Node version
+      placeholder: Codex CLI, macOS, Node 22 (do not include local paths)
+    validations:
+      required: true
+  - type: dropdown
+    id: install
+    attributes:
+      label: Installation path
+      options:
+        - One-command GitHub package
+        - Fresh clone and npm ci
+    validations:
+      required: true
+  - type: input
+    id: time
+    attributes:
+      label: Time to first successful demo
+      placeholder: 7m 30s, or did not complete in 15m
+    validations:
+      required: true
+  - type: textarea
+    id: explanation
+    attributes:
+      label: What does ThreadMesh do?
+      description: Explain it in your own words after using only the README.
+    validations:
+      required: true
+  - type: textarea
+    id: friction
+    attributes:
+      label: First confusing or failed step
+    validations:
+      required: true
+  - type: textarea
+    id: workflow
+    attributes:
+      label: Would you use it?
+      description: Name a real workflow, or explain why the capability is not useful.
+    validations:
+      required: true
+  - type: checkboxes
+    id: evidence
+    attributes:
+      label: Evidence checks
+      options:
+        - label: I found why delivery alone did not unlock the dependent task.
+          required: true
+        - label: I found what happened when the receiver was already running.
+          required: true
+        - label: The demo reported cleanup complete.
+          required: true
+        - label: I am independent of the ThreadMesh maintainer organization.
+          required: true
+        - label: I removed credentials, transcripts, IDs, and local paths.
+          required: true
@@ -19,10 +19,12 @@
 
 # ThreadMesh
 
-ThreadMesh is an experimental coordination protocol and JavaScript integration
-kit that lets one agent session notice an authorized dependency, decide whether
-to contact another session, and deliver a bounded suggestion without sharing
-global chat history or taking over the receiver.
+**Stop babysitting parallel coding agents.**
+
+ThreadMesh routes completion, blockers, review findings, and verified
+dependency state to the right agent session at a safe checkpoint—without making
+you copy results, spend model turns polling status, or let one session silently
+take over another.
 
 **The agent supplies the initiative. ThreadMesh supplies the boundary.**
 
@@ -31,12 +33,39 @@ global chat history or taking over the receiver.
 > release is suitable for local, trusted-process experiments—not production
 > authorization, hostile prompts, or multi-tenant deployment.
 
+## 76-second proof walkthrough
+
+<p align="center">
+  <a href="docs/assets/demo/threadmesh-proof-walkthrough.mp4">
+    <img src="docs/assets/demo/threadmesh-proof-walkthrough.gif" width="100%" alt="ThreadMesh evidence walkthrough: one kickoff, zero manual relay or polling, active receiver checkpoint, selective attention, and verified dependency unlock">
+  </a>
+</p>
+
+The walkthrough is generated from a fresh executable demo and retained real
+Codex evidence. It is not presented as a live screen recording. The local demo
+models the same four-handoff workflow two ways: the manual path requires at
+least one kickoff, four status checks, and four relay actions; the ThreadMesh
+path requires one kickoff and zero later relay or polling actions. Elapsed time
+and model tokens are deliberately marked **not measured** until a network-valid
+live baseline is retained.
+
+It also exercises the safety failure people worry about: when B is already
+running, the completion stays `pending` in a `checkpoint-offer`; B remains
+`running`, and the demo starts zero steer, interrupt, or native-turn operations.
+
+[Watch the MP4](docs/assets/demo/threadmesh-proof-walkthrough.mp4) ·
+[inspect asset provenance](docs/assets/demo/README.md) ·
+[run the proof yourself](docs/06-guides/attention-router-demo.md)
+
 ## Why this matters
 
-Running several agents in parallel creates a new coordination problem. Agent A
-may finish the exact input Agent B needs, but neither session knows when it is
-useful to speak. The user becomes a human message bus: notice the dependency,
-copy the result, find the right session, and explain why it matters.
+Running several agents in parallel often gives the user three extra jobs:
+
+- **clipboard:** notice A has the result B needs, then copy and explain it;
+- **poller:** repeatedly ask whether review, verification, or a dependency is
+  finished, consuming time and model quota even when nothing changed;
+- **traffic controller:** decide whether to queue, wake, steer, or interrupt B
+  without enough visibility into B's current work.
 
 ThreadMesh makes that handoff an explicit, portable capability:
 
@@ -47,9 +76,11 @@ ThreadMesh makes that handoff an explicit, portable capability:
 5. B's harness accepts, rejects, or defers it before model-context admission;
 6. the decision and delivery chain stays auditable.
 
-The intelligence is not “agents can send messages.” It is **selective
-initiative**: speaking when a dependency is real, staying quiet when it is not,
-and preserving the other session's agency.
+The intelligence is not “agents can send messages.” Transport is increasingly
+available from harness-native APIs, ACP, and A2A. ThreadMesh focuses on
+**selective initiative**: speak when a dependency is real, remain quiet when it
+is not, verify before unlocking downstream work, and preserve the receiver's
+agency.
 
 ## What proactive behavior looks like
 
@@ -110,12 +141,14 @@ five of five temporary sessions and all coordinator artifacts were removed.
 
 The completed result is deliberately classified `state=blocked` and
 `liveProductEvidence=false`: that retained run used fixture-owned or simulated
-Git and verification effects. The next branch now wires the existing bounded
-Git worktrees and process-isolated child verifier into the same correlated
-path. Its live rerun is pending after a reproducible local DNS/TLS failure; it
-has not been upgraded into product evidence. After that rerun, the remaining
-checkpoint is the manual relay/polling baseline and minimum critical
-negative/restart closure. Non-mainline expansion remains frozen.
+Git and verification effects. [#133](https://github.com/fyaic/threadmesh/pull/133)
+now binds the existing bounded Git worktrees and process-isolated child
+verifier into that correlated path on `main`. Its fresh real Codex rerun is
+pending after a reproducible local DNS/TLS endpoint failure; it has not been
+upgraded into product evidence. The deterministic manual-accounting baseline,
+active-receiver checkpoint negative, and 76-second evidence walkthrough are now
+the public product proof. New harness, transport, and generalized protocol work
+remains frozen until the live rerun and three external setup attempts close.
 
 [Read the exact fixture evidence](docs/09-reviews/2026-09-01-m5-2-autonomous-fixture.md) ·
 [Read the real Codex behavior](docs/09-reviews/2026-09-01-m5-2-real-codex-event-pump-behavior.md) ·
@@ -306,14 +339,14 @@ peer content or as a production security boundary.
   install Ajv and native `better-sqlite3`.
 - **Reference runtime:** authenticated JSON-RPC + SQLite coordinator for local,
   trusted-process experiments.
-- **Validation:** 379 tests, plus 55 schema cases and 7 transition cases;
+- **Validation:** 384 tests, plus 55 schema cases and 7 transition cases;
   documentation lint passes. These are separate counts, not one combined total.
 - **Default:** proactive coordination remains off unless a maintainer explicitly
   opts into the bounded experimental profile.
-- **Next mainline:** reuse the existing bounded Git-worktree and verifier
-  foundations inside the now-proven real Codex event-pump chain; add a manual
-  relay/polling baseline and minimum critical negative/restart evidence. Kimi
-  parity and broader hardening follow after M5.2 closure.
+- **Next mainline:** retain one network-valid real Codex traversal of the merged
+  real-effects path, run the measured manual baseline, and observe three
+  independent 15-minute setup attempts. Kimi parity and broader hardening
+  follow only after those product-proof gates.
 
 [Current status](docs/10-planning/project-status.md) ·
 [roadmap](ROADMAP.md) ·
@@ -325,8 +358,12 @@ peer content or as a production security boundary.
 | If you want to… | Start here |
 |---|---|
 | Understand the product | [What ThreadMesh is](docs/00-overview/product-guide.md) |
+| Watch the 76-second proof | [MP4 walkthrough](docs/assets/demo/threadmesh-proof-walkthrough.mp4) |
 | See real proactive behavior | [Real agent case portfolio](docs/06-guides/real-world-cases.md) |
 | Run the closed-loop local demo | [Attention-router demo](docs/06-guides/attention-router-demo.md) |
+| Audit the user-value baseline | [Manual relay/polling baseline](docs/06-guides/manual-relay-baseline.md) |
+| Audit non-interruption | [Active-session checkpoint case](docs/06-guides/non-interrupting-handoff.md) |
+| Try it as a new operator | [15-minute challenge](docs/06-guides/15-minute-operator-challenge.md) |
 | Compare selective model initiative | [End-to-end demo](docs/06-guides/end-to-end-demo.md) |
 | Add ThreadMesh to a harness | [Adapter implementation guide](docs/06-guides/implement-an-adapter.md) |
 | Evaluate a harness | [Harness support matrix](docs/00-overview/harness-support.md) |

@@ -19,21 +19,46 @@
 
 # ThreadMesh
 
-ThreadMesh 是一个实验性的 Agent 协调协议与 JavaScript 接入工具包。它让一个
-Agent session 在执行过程中发现经过授权的跨任务依赖，自主判断是否联系另一个
-session，并发送一条受约束的建议；同时不共享全局聊天记录，也不夺走接收方的
-上下文控制权。
+**不用再人工盯守多个并行 coding agent。**
+
+ThreadMesh 会把完成、阻塞、评审发现和经过验证的依赖状态，在安全 checkpoint
+路由给正确的 agent session：用户不用复制结果，不用消耗模型 turn 反复查询，也
+不会让一个 session 静默接管另一个 session 的当前工作。
 
 **Agent 提供主动性，ThreadMesh 提供边界。**
 
 > [!IMPORTANT]
 > ThreadMesh 目前是 pre-alpha，主动能力默认关闭。现阶段适合本地、可信进程范围
 > 的实验，不应作为生产级授权、多租户隔离或处理恶意 peer prompt 的安全边界。
 
+## 76 秒证据演示
+
+<p align="center">
+  <a href="docs/assets/demo/threadmesh-proof-walkthrough.mp4">
+    <img src="docs/assets/demo/threadmesh-proof-walkthrough.gif" width="100%" alt="ThreadMesh 证据演示：一次启动、零人工转发和轮询、活跃接收方 checkpoint、选择性注意力与验证后依赖解锁">
+  </a>
+</p>
+
+这个演示由一次新鲜的可执行 demo 和已经保留的真实 Codex 证据生成，不冒充实时
+录屏。对于同一个四次交接工作流，人工路径的最低成本是 1 次启动、4 次状态查询、
+4 次复制转发，共至少 9 次用户操作；ThreadMesh 路径是 1 次启动，后续 0 次转发、
+0 次轮询。耗时和 token 尚未实测，必须等网络正常的真实基线，文档不会虚构数字。
+
+演示还覆盖最重要的安全负例：B 正在运行时，完成事件只会以
+`checkpoint-offer` 留在 mailbox，decision 保持 `pending`；B 仍是 `running`，
+不会触发 steer、interrupt 或新的 native turn。
+
+[观看 MP4](docs/assets/demo/threadmesh-proof-walkthrough.mp4) ·
+[查看演示资产证据边界](docs/assets/demo/README.md) ·
+[亲自运行](docs/06-guides/attention-router-demo.md)
+
 ## 为什么需要它
 
-当多个 Agent 并行工作时，用户往往被迫充当“人工消息总线”：发现 A 的结果正好是
-B 缺少的输入，从 A 复制内容，找到正确的 B session，再解释这条信息为什么重要。
+当多个 Agent 并行工作时，用户往往被迫承担三份额外工作：
+
+- 当“剪贴板”：发现 A 的结果正好是 B 缺少的输入，再复制、查找、解释；
+- 当“轮询器”：不断询问评审、验证或依赖是否完成，即使状态没变化也消耗额度；
+- 当“交通警察”：在不了解 B 当前工作的情况下决定排队、唤醒、转向还是打断。
 
 ThreadMesh 把这个过程抽象成一项可移植能力：
 
@@ -44,8 +69,9 @@ ThreadMesh 把这个过程抽象成一项可移植能力：
 5. B 的 harness 在 checkpoint 接受、拒绝或延迟，再决定是否进入模型上下文；
 6. 完整的决策、投递与清理链路可审计。
 
-这里的“智能”不只是 Agent 会发消息，而是**有选择的主动性**：依赖确实存在时
-主动联系，无关时保持安静，并尊重另一个 session 的自主权。
+这里的“智能”不只是 Agent 会发消息。传输能力正在被 harness 原生 API、ACP 和
+A2A 普及；ThreadMesh 关注的是**有选择的主动性**：依赖确实存在时主动联系，
+无关时保持安静，验证后才解锁下游，并尊重接收 session 的自主权。
 
 ## 已验证的主动性效果
 
@@ -205,10 +231,10 @@ sandbox。不要用它处理任意恶意 peer 内容或充当生产安全边界
 - 协议：可执行 `0.0-draft`，仍可能调整。
 - 包：`@fyaic/threadmesh@0.1.0-alpha.0`，可从 GitHub 安装；根 export 是精简 SDK，CLI 与显式 runtime subpath 会安装 Ajv 和原生 `better-sqlite3`。
 - 参考 runtime：authenticated JSON-RPC + SQLite coordinator，面向本地可信进程实验。
-- 验证：378 项测试，加 55 个 schema case、7 个状态转换 case、文档与链接检查；这些计数分别报告。
+- 验证：384 项测试，加 55 个 schema case、7 个状态转换 case、文档检查；这些计数分别报告。
 - 默认策略：除非 maintainer 明确选择有边界实验 profile，否则主动协调保持关闭。
-- 当前边界：确定性 event pump 已能在一次 kickoff 后驱动 A→R→同一个 A→V→dependent，但三次真实 Codex event-pump 尝试分别停在产品探测、时间证据和用户暂停于五个 session bootstrap，尚未形成真实主动链；第三次没有执行正常信号清理，随后通过一次性精确操作删除并确认了五个测试 session 与临时资源。
-- 下一主线：冻结非主线扩展，先在现有实现上完成并保留一次真实 Codex event-pump 主动链；只有真实运行暴露出的 blocker 可以插队。随后再恢复真实 Git/verifier 闭环、Kimi parity 与可靠性矩阵。
+- 当前边界：第六次真实 Codex event-pump 已在一次 kickoff 后通过 9 个 native turn 完成 A→R→同一个 A→V→dependent，后续 runner phase prompt/direct activation 为 0，无关 session turn 为 0，清理 5/5；该次运行的 Git/verifier effect 是模拟的。真实 Git worktree 与 child verifier 已由 #133 合入同一路径，但组合后的新鲜 live 重跑仍受本机 DNS/TLS 故障阻塞。
+- 下一主线：在网络正常的 host 上保留一次真实 Codex real-effects 闭环，完成实测人工基线，并观察 3 位外部 operator 的 15 分钟上手过程。在这些产品证据前，继续冻结新 harness、transport 和泛化 protocol 扩展。
 
 [当前状态](docs/10-planning/project-status.md) · [路线图](ROADMAP.md) ·
 [协议草案](spec/README.md) · [验证记录](docs/09-reviews/README.md)
@@ -218,7 +244,11 @@ sandbox。不要用它处理任意恶意 peer 内容或充当生产安全边界
 - [中文文档入口](docs/zh-CN/README.md)
 - [英文文档总览](docs/README.md)
 - [产品说明](docs/00-overview/product-guide.md)
+- [76 秒演示](docs/assets/demo/threadmesh-proof-walkthrough.mp4)
 - [真实 Agent 案例](docs/06-guides/real-world-cases.md)
+- [人工转发与轮询基线](docs/06-guides/manual-relay-baseline.md)
+- [活跃 session 不打断案例](docs/06-guides/non-interrupting-handoff.md)
+- [15 分钟外部上手挑战](docs/06-guides/15-minute-operator-challenge.md)
 - [贡献指南](CONTRIBUTING.md)
 - [GitHub Discussions](https://github.com/fyaic/threadmesh/discussions)
 - [GitHub Issues](https://github.com/fyaic/threadmesh/issues)

@@ -150,13 +150,15 @@ nine bound native turns, zero later runner prompts or direct activations, an
 irrelevant zero-turn control, and exact cleanup.
 
 This exposed an execution-order imbalance rather than a change in product
-direction. The behavioral checkpoint is passed. The existing bounded
-Git-worktree and child-verifier foundations are now wired into that correlated
-path, but a successful live rerun is pending after a reproducible local DNS/TLS
-failure. The manual baseline and minimum critical negative/restart evidence
-remain next. New substrate, generalized recovery, cross-harness, or
-presentation work remains frozen. No partial integration attempt is promoted
-to M5.2 evidence.
+direction. The behavioral checkpoint is passed. The bounded Git worktrees and
+process-isolated child verifier are now wired into the correlated path by
+[#133](https://github.com/fyaic/threadmesh/pull/133), with deterministic
+positive and wrong-finding negative coverage. A successful live rerun is
+pending after a reproducible local DNS/TLS failure. The public demo now includes
+executable manual-action accounting and an active-receiver checkpoint negative;
+elapsed time and token cost remain unmeasured. New substrate, generalized
+recovery, cross-harness, transport, and protocol expansion remains frozen. No
+partial integration attempt is promoted to M5.2 evidence.
 
 - [x] Ship a one-command local demo with generated identities, grants, example
   sessions, and an inspector
@@ -186,16 +188,23 @@ to M5.2 evidence.
     and no new coordinator or verifier subsystem.
   - [ ] Retain one successful live Codex traversal of that real-effects path;
     current reruns are blocked by a reproducible local DNS/TLS endpoint error.
-  - [ ] M5.2 closure: reuse the existing bounded Git and verifier foundations
-    in one successful correlated run, add the manual baseline and minimum
-    critical negative/restart evidence, and keep raw product data out of public
+  - [x] Add executable manual workflow accounting: one kickoff plus four checks
+    plus four relays is a nine-action lower bound, versus one ThreadMesh kickoff.
+    Keep elapsed time and tokens explicitly unmeasured.
+  - [x] Add the active-receiver negative: a completion remains pending at a
+    checkpoint while B stays running, with zero steer, interrupt, or native-turn
+    starts.
+  - [ ] M5.2 closure: complete one successful correlated real-effects run and a
+    measured manual/live baseline while keeping raw product data out of public
     output.
   - [ ] M5.3: pass three fresh relevant runs plus the manual baseline,
     irrelevant, stale/unverified, restart, and cleanup matrix.
 - [ ] Repeat the loop across Codex and one ACP-compatible harness
   ([#93](https://github.com/fyaic/threadmesh/issues/93)).
 - [x] Publish the bounded inspector and reproducible deterministic evidence
   record ([#92](https://github.com/fyaic/threadmesh/issues/92)).
+- [x] Publish a 76-second evidence walkthrough generated from fresh executable
+  demo output, with retained real Codex evidence and honest claim boundaries.
 
 The executable closure gates for the real-agent phases are in the
 [M5 real Codex loop plan](docs/10-planning/m5-real-codex-loop.md). A local
@@ -209,6 +218,7 @@ incorrect dependency unlocks.
 ## M6 — Independent adoption and ecosystem bridges
 
 - [ ] Collect three independent setup attempts and one completed real workflow.
+- [x] Publish the 15-minute operator challenge and structured report template.
 - [ ] Close [#79](https://github.com/fyaic/threadmesh/issues/79) with independent
   harness-author feedback.
 - [ ] Make ACP the preferred multi-harness gateway.