From 9c368e983850e314aa54419971f5cfce1390e5db Mon Sep 17 00:00:00 2001 From: veil-chow-fyaic <247294299+veil-chow-fyaic@users.noreply.github.com> Date: Mon, 7 Sep 2026 20:24:41 +0800 Subject: [PATCH] docs: retain native readiness attempt without readable current-turn evidence --- ROADMAP.md | 4 +- docs/06-guides/codex-native-tasks.md | 5 ++ .../2026-09-07-native-public-entry.md | 67 +++++++++++++++---- docs/09-reviews/README.md | 2 +- .../community-followup-2026-09-07.md | 18 +++-- docs/10-planning/project-status.md | 7 +- docs/evidence/README.md | 1 + docs/zh-CN/codex-native-tasks.md | 4 ++ 8 files changed, 85 insertions(+), 23 deletions(-) diff --git a/ROADMAP.md b/ROADMAP.md index 42568db..c99f995 100644 --- a/ROADMAP.md +++ b/ROADMAP.md @@ -48,7 +48,9 @@ It needs no Node/MCP/hook setup. [One controlled opted-in desktop pair passed](d including prior context, B's own edit and busy/stop checks. A new [public-entry checkpoint](docs/09-reviews/2026-09-07-native-public-entry.md) verifies anonymous workflow retrieval; title-based activation, novice task selection and -normal plugin activation remain open. Sending is not race-free. External adapter/hook adoption is +normal plugin activation remain open. The live readiness attempt returned empty +current-turn evidence despite host-reported completion; resolve result visibility +before repeating it. Sending is not race-free. External adapter/hook adoption is a separate portability route, not a prerequisite for trying native guidance. The [desktop-first plan](docs/10-planning/desktop-entry-2026-09-07.md) supersedes diff --git a/docs/06-guides/codex-native-tasks.md b/docs/06-guides/codex-native-tasks.md index b92e79a..fbf2664 100644 --- a/docs/06-guides/codex-native-tasks.md +++ b/docs/06-guides/codex-native-tasks.md @@ -58,6 +58,11 @@ it. A missing workflow, missing tools or unresolved title is a stopped setup, not a successful connection. You can say **Stop ThreadMesh collaboration** in each task to cancel it; no separate control panel is required. +If the app shows “completed” but supplies no readable setup confirmation, keep +collaboration off. Do not assume pairing worked or repeatedly rerun it. This +occurred in our [readiness attempt](../09-reviews/2026-09-07-native-public-entry.md#authorized-live-readiness-attempt); +its cause is unresolved, and a completed indicator alone is not acceptance. + This is an explicit workflow request in the existing conversation, **not proof that installing a plugin hot-loads old tasks**. The model must actually retrieve the workflow, inspect available native tools and confirm the selected scope. diff --git a/docs/09-reviews/2026-09-07-native-public-entry.md b/docs/09-reviews/2026-09-07-native-public-entry.md index 242d220..7bb4ad5 100644 --- a/docs/09-reviews/2026-09-07-native-public-entry.md +++ b/docs/09-reviews/2026-09-07-native-public-entry.md @@ -1,7 +1,7 @@ # Native desktop public entry: retrieval and readiness boundary -Date: 2026-09-07. Status: **public retrieval checked; name-based desktop pairing -and activation not yet accepted**. This record does not replace the earlier +Date: 2026-09-07. Status: **public retrieval checked; live name-based readiness +attempt not accepted because current-turn evidence is empty**. This record does not replace the earlier [controlled native business case](2026-09-07-native-desktop-acceptance.md). ## What changed for a user @@ -28,7 +28,7 @@ avoids the inventory; no specific desktop reference-picker UI is asserted. | One measured fetch | 910 ms using Node fetch on the maintainer host | Cold-user timing or an onboarding speed guarantee | | Skill format validation | Passed the Skill Creator validator in an isolated `uv` dependency environment | Correct model decisions or plugin activation | | Repository regression | 459 passed, one optional native test skipped; 55 schema cases and seven transitions passed; 146 Markdown files linted cleanly | Live name resolution or autonomous model behavior | -| Existing native task name matching | Not run in this checkpoint | A title-only pairing pass | +| Existing native task name matching | Attempted in both tasks; completed-turn reads returned no items | Successful workflow retrieval, name matching or absence of peer sends | | Public-source activation and business edit | Not run in this checkpoint | End-to-end desktop onboarding | Public workflow revision: `592014782d10a8c4b88f46ea23b7cf588ff78355`. @@ -42,17 +42,60 @@ then passed using `uv run --with pyyaml python` and the existing validator. No user-global skill or plugin was installed. The published alpha.3 archive is unchanged; this revised workflow is reached through the pinned source URL. -## Next bounded check +## Authorized live readiness attempt -The earlier two dedicated native tasks remain stopped. A request was made for -permission to give each a readiness-only check and one task inventory, without -reactivating collaboration, sending peer advice or changing sample artifacts. -No such check has been dispatched at this checkpoint. +The user subsequently authorized one readiness-only message to each original +dedicated task and one task inventory per task. Both were `notLoaded` before +dispatch. The manager explicitly invoked these checks; this was not an automatic +peer wake or a race-free idle-send test. The messages provided the pinned public +URL and the other task's exact title, **not its internal ID**, and required a new +name lookup rather than reusing IDs remembered from the earlier run. -If authorized, record actual public retrieval, uniquely resolved target title, -available native tools, no outgoing peer message and unchanged business files. -Do not count already known IDs or a local workflow read as evidence for the -new entry. This is still a maintainer check, not an independent GUI-user result. +The narrow prompt required complete public retrieval, at most one authorized +inventory, actual native tool inspection, and a concise result. It explicitly +kept collaboration off: no peer messages, business edits, installation, permission +changes, new tasks or restoration of cancelled advice. The manager made exactly +two check dispatches and did not retry either model turn. + +| Observation | Task A | Task B | +|---|---|---| +| Host-reported terminal status | Completed, idle | Completed, idle | +| Host-reported duration | 142.592 s | 137.989 s | +| Current-turn items returned by native read | 0 | 0 | +| Current-turn result/tool evidence | Unavailable | Unavailable | +| Preceding turn still readable | Yes, six items | Yes, two items | + +A bounded reread including the preceding turn reproduced the empty-current-turn +result. No failure cause is established: a host-reported completion does not +prove the model ran the requested checks. Do not infer a quota failure, successful +name resolution, zero peer sends or a confirmed collaboration state from the +empty history. No enable command was issued; the last readable agreements remain +stopped. No private host storage or UI automation was used as a workaround. + +The three known sample artifacts were unchanged before/after by SHA-256: + +| Artifact | SHA-256, identical before and after | +|---|---| +| A's `brand.json` | `330315d837dd80782c73fb07ad641458a0259002f4fe0f98e191ccb67c9ed35e` | +| B's `landing.json` | `267643aca34e3da28f5807f6a49813545c7f849097bc1cf300cf3fee1870a3c7` | +| B's `price.txt` | `5f9ef93d379ed019937fefa3bada8830ad124e63342acc0aca6da0c0b6ffa38e` | + +Native read responses for the current and preceding turn are retained outside +the repository in owner-only local storage. Their file commitments are: +A `f8be6a2bc0ba08a8b15ff26190afa40ba39356f89856c3d3c4dbed950e197092`; +B `81443659b81d849e3cb2f998ec1e7411d5028681a760f04047bd06469715493f`. +No raw task IDs, histories or inventory contents were published. No screenshot +or recording exists for this attempt. The read responses are not full execution +traces and cannot fill the missing tool evidence. + +## Disposition + +The bounded readiness gate **did not pass**. The bilingual guide now explicitly +says that a completed indicator without a readable confirmation is not ready. +Resolve current-turn result visibility before another acceptance run; do not +paper over it with a new CLI pair, repeated quota-consuming dispatches or a +claim based on the earlier successful native case. Independent GUI onboarding +and public-source activation remain open. Official [skill documentation](https://learn.chatgpt.com/docs/build-skills) describes reusable instructions and host loading. It does not itself prove this diff --git a/docs/09-reviews/README.md b/docs/09-reviews/README.md index 36957d5..2af70ca 100644 --- a/docs/09-reviews/README.md +++ b/docs/09-reviews/README.md @@ -30,7 +30,7 @@ not count as live-product or independent external-verifier evidence. ## Live attempt audits -- [Native desktop public entry: retrieval checked, title pairing pending](2026-09-07-native-public-entry.md) +- [Native desktop public entry: retrieval checked, live readiness evidence unavailable](2026-09-07-native-public-entry.md) - [Codex-first installed-package acceptance and retained failures](2026-09-07-codex-first-use-release.md) - [Codex native desktop: prior context, original receiver edit and busy/stop checks](2026-09-07-native-desktop-acceptance.md) - [Codex native-task skill: earlier packaging/tabletop checkpoint](2026-09-07-codex-native-skill.md) diff --git a/docs/10-planning/community-followup-2026-09-07.md b/docs/10-planning/community-followup-2026-09-07.md index d14042c..35124bb 100644 --- a/docs/10-planning/community-followup-2026-09-07.md +++ b/docs/10-planning/community-followup-2026-09-07.md @@ -80,7 +80,7 @@ from the 273-second packaged example and removes the global Node badge: Node is a package requirement, not a requirement of the native skill workflow. Native attribution and the absence of measured incremental benefit remain explicit. -**Decision: draft only; do not post this turn.** The three requested maintainer +**Previous checkpoint decision: draft only.** The three requested maintainer example/progress/receiver-evidence improvements have shipped. Another README rearrangement is not a new functional release. The next useful increment is the remaining Codex desktop pairing/activation friction: test a publicly retrievable @@ -95,10 +95,14 @@ scheduled follow-up is configured by this document. Public-entry checkpoint: the bilingual title-and-topic prompt now uses a pinned public workflow; anonymous retrieval and byte parity passed. The readiness-only mode preserves previous stops. Title matching in the dedicated desktop pair has -not been dispatched pending scoped permission; do not claim the new entry has -passed merely because HTTP retrieval did. See the -[record](../09-reviews/2026-09-07-native-public-entry.md). The draft below remains -unsent, with no scheduled posting. +now been attempted with scoped user permission. Both host-reported completed, +but the current-turn native reads returned no items; preceding turns remained +readable, and sample artifact hashes were unchanged. This cannot prove title +matching or zero peer sends. See the +[record](../09-reviews/2026-09-07-native-public-entry.md). The bounded attempt is +complete, not passed. After this record merges, post the consolidated released +Codex example update with this explicit desktop limitation; do not delay the +contributor's already-delivered priorities behind repeated native retries. > Thank you, Andrei. We followed the three concrete priorities in your report. > @@ -137,8 +141,8 @@ advantage over native-only use. These are material clarifications to include whe a reply is authorized, not reasons to post another immediate acknowledgement. The user explicitly asked us to make several real improvements before replying -again. Therefore do not post another acknowledgement, roadmap promise, test -request or issue closure now. Continue implementation, review and verification +again. Do not post another acknowledgement-only message, roadmap promise, test +request or issue closure. Continue implementation, review and verification inside the authorized task rather than stopping after each small step. The next reply should contain the actual released install/run entry, a short diff --git a/docs/10-planning/project-status.md b/docs/10-planning/project-status.md index 642c216..8b7073d 100644 --- a/docs/10-planning/project-status.md +++ b/docs/10-planning/project-status.md @@ -6,8 +6,11 @@ title-and-topic prompt with a pinned public workflow. Anonymous retrieval and source parity passed; readiness checking is separate from activation. The [entry record](../09-reviews/2026-09-07-native-public-entry.md) keeps title-only -desktop matching and end-to-end activation open. The earlier test pair remains -stopped; no peer probe or reactivation was dispatched for this checkpoint. +desktop matching and end-to-end activation open. Two user-authorized read-only +checks were dispatched: both host-reported completed, but current-turn histories +were empty while preceding turns were readable. Sample hashes were unchanged; +readiness and absence of peer sends cannot be verified. No enable command or +repeat dispatch was issued. Resolve result visibility before another live gate. **Value and evidence correction:** the native demo uses Codex's own communication and continuation; the skill adds guidance, not transport. [English responsibility diff --git a/docs/evidence/README.md b/docs/evidence/README.md index d094fa0..87f37ee 100644 --- a/docs/evidence/README.md +++ b/docs/evidence/README.md @@ -6,6 +6,7 @@ adoption or proof of incremental value over the host's native capability. | Recent case | Public record | Retained private originals | |---|---|---| | Codex desktop prior-context pair | [Actual exchange, B's diff and reduced audit](codex-native-2026-09-07/README.md) | Complete A/B native exports, original tasks and files | +| Codex public-entry readiness attempt, not accepted | [Empty current-turn evidence and unchanged artifacts](../09-reviews/2026-09-07-native-public-entry.md) | Native read responses retained separately; not full execution traces | | Codex default installed-package pass, 272.604 s | [Acceptance](../09-reviews/2026-09-07-codex-first-use-release.md) | Events, report, transcripts and sample artifacts copied to a dated non-temporary archive | | Codex diagnostic pass, 184.050 s | [Same record, separate budget](../09-reviews/2026-09-07-codex-first-use-release.md) | Same archive categories; not counted as the default gate | | Codex old-runtime failure | [Failure retained alongside passes](../09-reviews/2026-09-07-codex-first-use-release.md#failures-that-changed-the-implementation) | Events and failed-run report copied to the dated archive | diff --git a/docs/zh-CN/codex-native-tasks.md b/docs/zh-CN/codex-native-tasks.md index 05ff42c..6b0fec1 100644 --- a/docs/zh-CN/codex-native-tasks.md +++ b/docs/zh-CN/codex-native-tasks.md @@ -47,6 +47,10 @@ https://raw.githubusercontent.com/fyaic/threadmesh/592014782d10a8c4b88f46ea23b7c 流程读不到、工具缺失、名称无法确定，都属于设置未完成，不能当作连接成功。 要取消，可以分别在两个任务里说“停止 ThreadMesh 协作”，不需要额外控制面板。 +如果应用显示“已完成”，却没有可读的设置确认，请保持协作关闭，不要当作配对成功， +也不要反复重跑。这是我们在[只读就绪检查](../09-reviews/2026-09-07-native-public-entry.md#authorized-live-readiness-attempt) +中实际遇到的问题，原因尚未确定；“已完成”的状态不能代替验收结果。 + 这是在已有对话中明确请求使用流程，**不是证明安装插件能热加载旧任务**。 模型应实际读取流程，核对已有原生工具，并确认选择范围。若读取失败或缺少工具， 就停在能力检查，不能用新 CLI 会话或私有接口冒充成功。