dsh-global-todo 0.16.3 → 0.17.1

package/CHANGELOG.md

@@ -1,3 +1,47 @@
+## [0.17.1] — 2026-10-24
+
+Two READMEs instead of one: Chinese in `README.md`, English in `README.en.md`, and the switch to
+the other language on the first line of each.
+
+- **Two READMEs, one language each.** `README.md` is the Chinese document — the name GitHub's
+  repository page shows and the only one npm renders — and `README.en.md` is the English one; each
+  opens with a one-line switch to the other. The 0.17.0 npm page had an English document with the
+  Chinese only reachable through a relative link, and relative links are exactly what npm rewrites
+  or drops, so the switch is an **absolute** GitHub URL (a CJK heading slug is not stable across
+  renderers either). `README.zh.md` becomes a two-line landing page so old links still resolve;
+  `README.en.md` joined `files`, and `test/manifest.test.mjs` now reads the Node and DSH floors out
+  of the two real documents. AGENTS.md records the convention.
+
+## [0.17.0] — 2026-10-24
+
+Documentation and prompt text: the confirmation gate is the headline now, not the third bullet.
+Nothing in the write path changed — no field, no route, no validation rule — so upgrading is safe
+for existing stores.
+
+- **A release at all, because npm freezes documentation at publish time.** The README, `docs/` and
+  the plugin-page locales in the tarball are a snapshot; leading with the gate only reaches the npm
+  page and the installed docs through a version.
+
+- **Both READMEs lead with the gate.** It is claim 01 (was 03) and it names **both doors**: an
+  agent's `global_todo_done` (or a status change to `done`), and the host closing a dispatched todo
+  after a clean turn. A new early section — 「完成确认：没有一条记录能自己变成「完成」」 / *Confirmation:
+  no record finishes itself* — carries the two-door table, the two answers and the clean-turn rule;
+  the old 0.7 subsection and its row are gone rather than duplicated, and the comparison table gained
+  a *Finishing* row.
+- **The landing page, the npm description and the plugin-page locales open with the same sentence**:
+  page title, `meta`/`og`/`twitter` descriptions, hero lede, claim 01 (was 03), and the compare
+  table. `package.json` gained `human-in-the-loop`, `confirmation` and `approval`, which is how
+  someone looking for this actually searches.
+- **`docs/DESIGN.md` documents `review` for the first time.** It lived in `lib/model.js` and in no
+  design doc: there is now a *Done is a claim, from either door* section, a `review` field row, and
+  the `links` row finally admits `kind: 'image'` (0.15 shipped it).
+- **The model-facing half says it too.** The `global_todo_done` and `global_todo_update`
+  descriptions call a completion a claim; the done result prints 「这条挂在「待确认」…」; `todoLine()`
+  marks pending records in every rendering, so the model cannot read one as settled; and the injected
+  guidance tells the model to reply 「已标记完成，等你确认」 instead of 「已经结束了」. The stale
+  「插件也会在你安静下来后自动划掉」 line is gone — that idle-window heuristic was removed in 0.3.0
+  and the sentence outlived it by thirteen minor versions, telling the model the opposite of the rule.
+
 ## [0.16.3] — 2026-10-23
 
 Documentation release: the npm page finally shows the current install instructions.

package/README.en.md

@@ -0,0 +1,608 @@
+**English** · [中文](https://github.com/LUgABeEEE/dsh-global-todo/blob/main/README.md)
+
+# dsh-global-todo
+
+<img src="docs/icon.svg" width="52" height="52" alt="全局待办 icon: a list card with a check and an arrow leaving it">
+
+**A cross-session todo list for DeepSeek Harness: one file you own, written by you and your
+agent, and handed to a conversation in one click.**
+
+Three things it does that a session-local plan cannot — the first is the one it is known for:
+
+1. **An agent saying "done" is a claim, whichever door it comes through.** Its own report — the
+   tool call, a status change — and the host closing the todo after a clean turn are both just
+   claims: the record waits in **待确认 / Needs review** until a person confirms it, or sends it
+   back with its completion cleared. A session-local plan never asks that question: the agent
+   rewrites the table and the work is "finished".
+2. **One JSON file is the whole list.** `cat`, diff, commit it, put it on a NAS, carry it to
+   another machine. Nothing lives inside a session log.
+3. **Hand-off is a button, not a paragraph.** Four modes, the working directory checked, and
+   the todo's own **steps** arrive as a numbered plan.
+
+![The panel in use: list, detail, steps, review, inbox, reminders, selection](docs/screenshots/tour.gif)
+
+```sh
+dsh plugin add dsh-global-todo
+```
+
+[Live page](https://lugabeeee.github.io/dsh-global-todo/) (22-second typographic demo) · [中文](https://github.com/LUgABeEEE/dsh-global-todo/blob/main/README.md#zh) · [5-minute quickstart](docs/QUICKSTART.md) · [FAQ](docs/FAQ.md) · [Every state has a baseline image](#screenshots) · [Design notes](docs/DESIGN.md)
+
+---
+
+## Why another todo plugin
+
+DSH's built-in `todo_write` belongs to one agent session: whole-list replacement, three
+states, and a plan that disappears when the session does. The community plugins that fix
+that split into two camps — the ones that dispatch work to agents (auto-resume, bound
+sessions) and the ones that accept real document attachments. This plugin does both, on
+one record:
+
+| | built-in `todo_write` | dsh-global-todo |
+|---|---|---|
+| Owner | one agent session | **you** — across sessions and workspaces |
+| Writing | whole-list replacement | per-field, human **and** agent |
+| When it was added | not tracked | **`createdAt`, immutable, shown everywhere** |
+| Context | title + state | description + **documents and addresses** |
+| Hand-off | — | **one click: reuse the bound conversation or open a session** |
+| Finishing | whatever the agent writes | **waits in Needs review for a person** |
+| Storage | session event log | one JSON file you can `cat`, edit, diff and back up |
+
+## Confirmation: no record finishes itself
+
+This is the rule the plugin is known for, and the one that separates it from every list that lets
+the agent report its own progress: **an agent's "done" is a sentence it said.** The record is
+written as `done` and carries `review: { pending: true }`; the row shows a **Needs review** badge
+and the header a chip with the count, until a person answers.
+
+Both doors are gated, because both are only a claim:
+
+| Where "done" came from | The record | Who answers |
+|---|---|---|
+| the agent's `global_todo_done`, or a status change to `done` | `done` **+ pending review** | a person |
+| the dispatched conversation ended cleanly, so the host closed it (`turn/end` = `completed`) | `done` **+ pending review** | a person |
+| you clicked done in the panel | `done` | already a person |
+
+- **Two answers.** *Confirm done* drops the mark and records who confirmed it; *Send back* returns
+  the todo to `open`, clears `doneAt` and the completion record, and adds a `reopened` entry.
+  Nothing is deleted: title, description, steps, attachments and notes all stay.
+- **The automatic close goes through the same gate.** What the host writes on the agent's behalf is
+  thinner than what the agent could have written — only the closing message and the files that turn
+  changed — so it has even less business counting by itself.
+- **Only clean turns are candidates.** `error`, `aborted`, `interrupted` and `max-tokens` leave the
+  todo open. A failed job wearing a "done" label is the one failure mode that would make the whole
+  list untrustworthy.
+
+There is exactly one implementation of the rule: `completeTodo`, `applyPatch` and a status change in
+`lib/model.js` share the same `by === 'agent'` test, and the panel and the ten tools all go through
+it.
+
+## What 0.9 and 0.10 added (and the two just before)
+
+Five surfaces these READMEs had never explained. Session read-back and evidence came in 0.9;
+Markdown and the weekly report in 0.10; steps in 0.8. Each changes what a record can carry, or what
+you can do with it after the fact. (The review mark came in 0.7 and now opens this document, instead
+of hiding among the new features.)
+
+### Steps: a plan inside the todo
+
+A todo can carry up to 20 steps of 200 characters (`LIMITS` in `lib/model.js`). A step is
+`{ text, done }` — a line inside one unit of work, not a subtask with its own id and status.
+
+- **Add them** in "More fields" → Steps, one per line.
+- **Tick them** in the detail panel; the row shows `steps 2/5`.
+- **Hand off** and the dispatch prompt numbers them, prefixing finished ones with `[已完成]`.
+- **Export** and Markdown writes them as nested `- [ ]` lines; import reads them back.
+
+### Evidence: the sentence it came from
+
+`evidence` is the sentence a captured todo came from, quoted verbatim, up to 300 characters.
+`global_todo_capture` asks the model for it, `global_todo_add` accepts it, and the detail panel
+shows it as a blockquote under **In the author's words**. A todo you typed yourself has none.
+
+The field exists for one reason: a capture without its source reads as an opinion later. It is
+also in the Markdown export (`原话:`).
+
+### Session read-back
+
+A hand-off starts work in another conversation, so the detail panel answers "what did it say"
+without switching away. `GET /api/global-todo/session?id=` returns `alive`, `dir` and a `tail`
+of the last assistant message; the panel shows **Session read-back** with running / dormant, the
+last sentence, and a button that opens that conversation.
+
+The tail is bounded and in memory: 200 conversations, 400 characters each, filled from every
+`assistant/message`. Nothing is written to disk, so a Host restart forgets it. A missing tail is
+normal — this process has not seen that conversation speak.
+
+### Markdown in and out
+
+The JSON export is the round-trip format. Markdown is the one you read, diff and edit by hand
+(note that `doing` and `cancelled` come back as `open`: a hand-edited list has no work-in-progress
+state, and the JSON remains the format that round-trips exactly):
+
+```text
+- [ ] 标题
+      优先级: 高 · 截止: 2026-10-21T18:00 · 标签: 发布, 文档
+      原话: …
+      描述: 第一行
+      - [x] 一步
+```
+
+| Direction | Where | Behaviour |
+|---|---|---|
+| Out | **More actions → Export Markdown**, or `GET /export?format=md` | Grouped 未完成 / 已完成 / 已归档; metadata indented under each task |
+| In | **More actions → Import** (a `.md` file), or `POST /import { markdown }` | Creates new todos through `create`; Markdown has no ids, so it never overwrites an existing record |
+
+Tolerant by design: unknown lines are ignored, an unparseable date is dropped, and a file with no
+`- [ ]` line is refused with a reason. Field labels are Chinese (`优先级 / 截止 / 标签 / 原话 /
+描述 / 附件 / #`), and import reads the same set.
+
+### Weekly report
+
+**More actions → This week** (`GET /report?days=7`) answers "what moved, and did any of it land
+in code". Everything is derived from completion records; nothing is stored.
+
+| Field | Meaning |
+|---|---|
+| `finished` | completions in the window, newest first, with the actor and the summary |
+| `sessions` | distinct conversations those completions came from |
+| `created` | todos added in the window |
+| `pendingReview` | done items still waiting for an answer |
+| `files` | file artifacts touched, most frequent first — 12 listed, `fileCount` total |
+
+`days` above 90 is clamped to 90; a missing or non-positive value falls back to 7.
+
+## Screenshots
+
+![The panel: count chips, grouped list, urgent rows marked](docs/screenshots/panel.png)
+
+![One row opened: inline editing, attachments with 打开 / 复制路径, activity](docs/screenshots/panel-detail.png)
+
+![Settings: grouped fields, hints under each label, one save row](docs/screenshots/panel-settings.png)
+
+Every state that only appears under a condition has a fixed baseline image, rendered by the same
+preview and photographed by the same tool:
+
+| State | Image |
+|---|---|
+| Empty list | [`panel-empty.png`](docs/screenshots/panel-empty.png) |
+| Filter with no match | [`panel-filtered.png`](docs/screenshots/panel-filtered.png) |
+| Inbox (awaiting confirmation) | [`panel-inbox.png`](docs/screenshots/panel-inbox.png) |
+| Awaiting review (agent finished) | [`panel-review.png`](docs/screenshots/panel-review.png) |
+| Due reminder | [`panel-reminders.png`](docs/screenshots/panel-reminders.png) |
+| Selection mode | [`panel-select.png`](docs/screenshots/panel-select.png) |
+
+These are produced by `npm run screenshots` (which chains `npm run preview`): the real client
+half is rendered to static HTML with no Host and no dependencies, then photographed at 1280×720
+into `docs/screenshots/`. Playwright is not a dependency — the tool borrows the copy a DSH
+profile already installed, or whatever `DSH_PLAYWRIGHT` points at.
+
+What each file shows, and when it is regenerated, is indexed in
+[docs/screenshots/README.md](docs/screenshots/README.md).
+
+## Requirements
+
+- DeepSeek Harness `0.2.x` (developed and verified against `0.2.1-rc.2`; `engines.dsh` is
+  `>=0.2.0-rc.1 <0.3.0`, and a Host outside it skips the bundle instead of warning)
+- Node.js `>= 22.19.0` (the Web profile's own requirement)
+- A **Web** client: the panel is served by the Host's web server. Headless / ACP / SDK keep the
+  agent tools but have no panel; an Electron-style desktop client (file:// + IPC) is not
+  supported yet
+- After a Host upgrade, refresh the page — the client module graph belongs to the release the
+  page booted with
+- No runtime dependencies. Nothing is imported from `@deepseek-ai/*`.
+
+## Install
+
+**Check your profile name first.** The plugin is installed into one profile, not into the
+machine; a wrong profile name still exits 0, and then nothing shows up in the sidebar after a
+restart — the most common first-run trap.
+
+```sh
+ls ~/.dsh/profiles          # Windows: dir %USERPROFILE%\.dsh\profiles
+```
+
+Replace `<profile>` below with one of the names that command prints (`web` on most hosts;
+use your desktop or `headless` profile name if that is what you run):
+
+```sh
+# from npm
+dsh plugin --profile <profile> add dsh-global-todo
+
+# or from a checkout, which is also the development path
+dsh plugin --profile <profile> add /absolute/path/to/dsh-global-todo
+```
+
+The plugin is a Loader row, not a hot-loaded script: **restart the host** (`dsh web`, or
+whatever start command you use) after installing, then refresh the page. A "待办 / Todos"
+entry appears in the sidebar.
+
+> Sidebar still empty? Check three things in order: ① which profiles `ls` actually printed and
+> whether the one you installed into is the one running; ② whether the host really restarted
+> (the Loader is read once at boot); ③ whether the page was refreshed (the client module graph
+> belongs to the release the page booted with).
+
+The npm package is the source of truth. A community catalog can lag behind it or point at an
+older release; pin the version you want (`dsh plugin --profile <profile> add dsh-global-todo@0.13.1`)
+when it matters.
+
+### Installing from DSH's Discover tab
+
+DSH's Discover tab reads a plugin market index from a URL you add yourself — dsh ships none, and
+a user-added index is the one path that installs without copying a command. This repository
+publishes one:
+
+```text
+https://lugabeeee.github.io/dsh-global-todo/market/index.json
+```
+
+It is generated from npm by `npm run market` (`docs/market/index.json`) and validated against
+dsh's own `@deepseek-ai/dsh-plugin-market-index` before it is written. `npm test` fails when its
+newest version is not the version in `package.json`, so a release that forgets to regenerate it
+is caught before it ships.
+
+**Hand-copied a directory instead?** Check it before wondering why the plugins page looks wrong:
+
+```sh
+npx dsh-global-todo-doctor "$HOME/.dsh/plugins-local/全局待办插件"
+```
+
+It verifies the manifest, the patch file, both entries, `locale/`, the icon and `lib/`, and exits
+non-zero when something required is missing — the `locale/` case is the field failure that makes
+the desktop app print `Plugin metadata … ENOENT … \locale` while the plugin still loads.
+
+To roll back:
+
+```sh
+dsh plugin --profile <profile> remove dsh-global-todo
+```
+
+Your list is not touched by install or uninstall: it lives outside the plugin, at
+`$DSH_HOME/global-todo/todos.json`.
+
+## Using it
+
+### The panel
+
+- **Add** — type a title and press Enter. "More fields" reveals the description, the
+  attachment box (one document path or URL per line), tags, priority and a due time. An
+  attachment may also be `{ kind: "image", value: "/abs/shot.png" }` — the row then shows a
+  thumbnail and the detail a preview, with the bytes read on demand (png/jpg/jpeg/webp/gif/avif,
+  ≤8 MB) and never copied into the list.
+- **Filters** — all / open / doing / done / overdue, plus a text search across titles,
+  descriptions, tags and attachments. The count chips in the header are filters too.
+- **Grouped by urgency** — overdue / today / next up / later / no due date / done, each
+  collapsible, because "what is late" is the first question a list has to answer.
+- **Rows** — show `added <time> (<age>)` with the exact timestamp on hover, the priority,
+  the due time, the tags, the attachment count, and which side created the entry.
+- **Details** — edit title, description, priority and due time inline; attach or detach
+  documents and addresses (with 打开 / 复制路径 on each); read the activity trail.
+- **Undo a delete** — deleting keeps the record for eight seconds and offers 撤销, which
+  restores it verbatim: same id, same creation time, same attachments.
+- **Keyboard** — `/` focuses search, `n` focuses the new-todo field while the add form is
+  open, `Esc` collapses the open detail. Row actions appear on hover or keyboard focus.
+- **Reachable from the LAN and from a tunnel** — the routes defer to the Host's connection fence
+  (which honours `--trusted-host`); the plugin's own loopback check is only the fallback for a
+  Host without that service. `/health` reports `fence` so you can see which one is in force.
+- **Auto-collect (off by default)** — a panel switch. When on, a conversation you are in that
+  finishes a clean turn is asked once whether anything needs following up, but only once it has
+  been quiet for 30 minutes, and only if you are actually in that conversation (never a dispatch
+  target, a subagent or a background run). Aborted and failed turns are skipped. The question is
+  itself a message in the conversation, so it restarts the quiet clock — which is why **only a
+  message you wrote arms it again**: one question per turn from you, and it can never chase its own
+  answer. The switch is Host state, so it survives the panel being closed and is reported by
+  `/health`. Each conversation's group also carries an **`Exempt`** button: it takes that
+  conversation's whole **working directory** out of automatic collect (click again to undo). That
+  is for conversations a machine drives — a chat companion's room — whose session is recreated with
+  a new id when it restarts, so the exemption is keyed on the directory rather than on one
+  conversation and cannot quietly stop applying.
+- **Collect per conversation** — the `Collect` action sits on each conversation's group header in
+  the list, asks *that* conversation to record what needs following up, and the items land under
+  it. The toolbar keeps a "collect from this chat" fallback, shown only when the current
+  conversation has no group yet. A conversation that is not running is resumed first, like a
+  hand-off.
+- **Batch assignment** — switch on `选择`, tick a few todos, and one click sends each one to the
+  conversation that wrote it, else to a same-workspace conversation, else to a target you pick.
+  That is an attempt order: an unloadable target means the next one is tried. Anything with
+  nowhere to go comes back with a picker rather than being sent somewhere unchosen, and one
+  failure does not abort the batch.
+- **Search syntax** — `tag:release` (or `#ops`), `is:overdue`, `is:active`, `has:link`, `before:2026-11-01` / `after:2026-10-01` (deadline bounds; a half-typed date simply stops filtering),
+  `has:note`, `level:p0`, plus free words combined with AND (which also match note text). A
+  search without a colon behaves exactly as it always did.
+- **Progress notes** — an append-only trail per todo (count on the row, full text and an add
+  box in the detail). An agent reporting a blocker writes here instead of rewriting your
+  description, and writing one makes automatic completion stand down.
+- **Dead attachments are flagged** — a file link whose target is gone is marked on the row and
+  labelled in the detail. Computed on read, never stored.
+- **Where a new conversation goes** — a hand-off that opens one joins the workspace owning
+  the conversation that wrote the todo (and starts in that workspace's directory), so
+  "continue this work" does not land beside unrelated sessions. The picker also lists the
+  workspaces (`＋ new conversation · <workspace>`), and a named workspace wins over the
+  inherited one. The working directory only decides when neither is known.
+- **The picker lists what you actually use, newest first, without archived conversations.** The
+  order is the sidebar's recency key (`updatedAt`). Conversations sharing a title get a short id
+  suffix, and a target that is no longer listed (archived since it was chosen) is shown as
+  `(archived)` instead of appearing as a different conversation.
+- **You pick the target** — the toolbar shows `交给：[ conversation ▾ ]`, listing your
+  conversations: whichever you choose is sent with `authoritative: true` and the Host honours
+  it verbatim, with no inference on top. The list marks the conversation the Host would have
+  chosen on its own, `自动` restores that behaviour, and `＋ a new conversation` starts a fresh
+  context. A choice is remembered, and changing it is one click in the same control — what you
+  see is what happens.
+- **A receipt for every hand-off** — the panel says which conversation received the work and
+  offers to open it. A dispatch starts work *elsewhere*, so staying silent looked exactly
+  like "nothing happened".
+- **Hand to agent** — the panel says **where it will go** before you press it, and the
+  default is `current`: the conversation **you** last wrote in (read from the session log,
+  keeping only `source.kind === 'user'` messages — the plugin's own injected hand-off
+  messages are excluded, or each dispatch would drag the next one to its own target; the id
+  is also kept next to the data so a restart does not forget it). `auto` prefers the
+  conversation this todo already owns, then the current one, and only then opens a fresh
+  session; `new` always opens one; `bound` insists on the bound conversation and says why
+  if it is gone. When nothing usable can be resolved the hand-off **fails with a reason**
+  instead of quietly opening another conversation.
+- **Grouping** — the toolbar offers `conversation / time / level` and defaults to
+  **conversation**: everything one agent conversation submitted collapses into a single
+  entry that says how many it holds, so a few 「抓一下待办」 runs do not turn the list into a
+  wall of rows. A conversation with one item shows that row directly; items with no source
+  conversation (added in the panel) sit under 手动添加. `time` is the original
+  overdue / today / soon / later / no-due buckets, `level` groups by priority.
+- **Levels P0–P3** — `P0 urgent / P1 high / P2 normal (default) / P3 low`; the API accepts
+  either the canonical names or the labels (`priority: "P0"` works), with a badge on the
+  row. Records written before levels existed keep their `high/normal/low`, so nothing needed a
+  migration. A conversation group shows the highest level it holds, and a green dot marks a
+  group with a conversation still running.
+- **Collect from this chat** — one button injects "record what this conversation needs to
+  follow up, with `global_todo_capture`" into the conversation you are in, so a chat turns
+  into todos without typing them out. The prompt states what *not* to record as well
+  (session plans, small talk, finished work) — a capture instruction without that ceiling
+  only produces noise.
+- **Automatic completion** — when a handed-off session closes its turn **cleanly**, the Host
+  finishes the todo itself: the agent's closing message becomes the summary and the files
+  that turn touched become the artifacts. No waiting for the model to remember a tool call.
+  A finished row shows 完成于 … plus artifact chips, and its detail carries a completion
+  section you can open or copy paths from. Two guards matter: only a session this plugin
+  **just dispatched to** can complete itself (idle chatter in a bound conversation must not
+  finish work), and a turn that **failed or was interrupted is never marked done** — an open
+  todo is far better than a failed job wearing a "done" label. `global_todo_done` with
+  `summary` and `artifacts` remains the primary path; the watcher is the safety net.
+- **Working directory** — the toolbar says where a new session will start. It follows the
+  conversation you are in by default; 选择 picks one explicitly and remembers it, and
+  跟随当前会话 clears that choice. A todo also remembers the directory it was written
+  from (taken from the tool call's own context), and that is inherited too. With none of
+  the three available, the hand-off is **refused with a reason** — a session without a
+  working directory is a session where the agent can do nothing.
+
+### Agent tools
+
+| Tool | Purpose |
+|---|---|
+| `global_todo_add` | Record an item — description, attachments, steps, tags, priority, due time |
+| `global_todo_note` | Append a progress note — where you are, what blocks you, who must decide |
+| `global_todo_capture` | **Record a batch**: for 「抓一下待办」 or a turn that produced several follow-ups; each entry is validated on its own, and entries scoring below `captureMinWeight` are reported as skipped with their weight instead of being registered |
+| `global_todo_list` | Query by state, tag or text; every row carries its creation time |
+| `global_todo_get` | Read one item in full, including attachments and the last hand-off |
+| `global_todo_update` | Change any field of one item, including `status` (`archived` archives without deleting); `""` clears, `[]` empties |
+| `global_todo_link` | Attach or detach documents and addresses |
+| `global_todo_done` | Mark done, back to open, or in progress |
+| `global_todo_reopen` | Put a **wrongly finished** item back (`done` / `cancelled` → `open`); refuses anything else, and the trail records it as the agent |
+| `global_todo_remove` | Delete permanently (done items should be marked, not deleted) |
+
+### Live check
+
+The test suite uses doubles; `npm run check:live` checks the **wired** Host instead —
+the routes the panel actually calls, the list it actually reads, and a record that survives
+delete-then-undo. It creates a temporary record, compares every field after the round trip
+(id, creation time, attachments, tags, level, due time, description), asserts that bound
+records carry `sessionAlive`/`sessionDir`, and removes its temporary record even when it
+fails. It needs a running Host, so it is not part of `verify`.
+
+### Proactive capture
+
+A guidance section is injected into the system prompt (turn it off with `guidance: false`).
+It states both sides of the policy, because a prompt that only says "remember things"
+produces noise, and a noisy list is worse than an empty one:
+
+- **record on your own initiative** when the user mentions something that needs following
+  up across sessions (waiting on a reply, due next week, a decision pending, a promise),
+  when a turn ends with a step that should not continue in that session, or when the user
+  says 「记一下 / 别忘了 / 回头办」;
+- **do not record** what you can finish in this step, the session's own plan, or small talk;
+- say in **one line** what you recorded — do not restate the description;
+- say **「抓一下待办」** and the agent scans the recent conversation, records the batch with
+  `global_todo_capture` and answers with a summary.
+
+The boundary: it **records, it never starts work**. The list belongs to the human, and
+starting a task stays a human decision (see the non-goals in [ROADMAP](docs/ROADMAP.md)).
+
+Capture is not allowed to become noise. Before an item is written, the same normalized title
+(full-width, case, spacing and punctuation folded away), a close paraphrase within the last 72
+hours (Dice similarity over character bigrams, 0.8), and the same errand finished in the last 7
+days are all refused with a reason on the record; archived and cancelled rows do not block a
+new one. The rules live in `lib/capture.js` and are tested on their own.
+
+### HTTP API
+
+The panel talks to the host over these routes. They are fenced by the host connection
+fence when that service is present, and by a strict loopback check otherwise.
+
+| Method | Path | Body / query |
+|---|---|---|
+| GET | `/api/global-todo/list` | `status`, `query`, `tag`, `sort`, `limit` — the reply carries `matched` and `truncated` when the page was cut |
+| GET | `/api/global-todo/dirs` | `path?` — list subdirectories for the panel's built-in folder picker (the shell's own picker is preferred when mounted) |
+| GET | `/api/global-todo/get` | `id` |
+| GET | `/api/global-todo/export` | — (the store's own JSON document; `?format=md` gives a Markdown task list) |
+| POST | `/api/global-todo/import` | `{ doc, mode?: replace\|merge, dryRun? }` — `replace` is an exact round trip, `merge` updates by id and appends (what the panel sends); `dryRun` returns `wouldCreate` / `wouldUpdate` and writes nothing |
+| POST | `/api/global-todo/reminders/ack` | `{ ids }` — stop listing these passed deadlines as due |
+| POST | `/api/global-todo/note` | `{ id, text }` — append a progress note |
+| POST | `/api/global-todo/capture` | `{ sessionId? }` — ask the current conversation to collect its follow-ups |
+| POST | `/api/global-todo/bulk` | `{ ids, status?, addTags?, removeTags? }` — one change across a selection; a stale id lands in `unmatched` instead of failing the batch |
+| POST | `/api/global-todo/ingest` | `{ items: [{ title, source: { kind }, confidence }], dryRun? }` with `Authorization: Bearer <settings.ingestToken>` — external items; below `ingestConfirmBelow` they wait in the inbox |
+| GET | `/api/global-todo/health` | — (store location, counts, dispatch capabilities, plus **tool registration** and **whether the completion watcher attached** — both fail quietly, so both are reported — and `metrics`: dispatch success rate, agent tool calls, list growth; `autoCollect`: why each waiting conversation was or was not asked, counters, and how the settings file loaded; `version` is the running module and `packageVersion`/`stale` say whether the installed package is newer than the running one; `review.items` names the finished todos that still wait for a person, so a check does not have to drive the panel) |
+| POST | `/api/global-todo/create` | `{ title, description?, links?, steps?, tags?, priority?, dueAt? }` |
+| POST | `/api/global-todo/update` | `{ id, ...fields, by?: human\|agent }` |
+| POST | `/api/global-todo/status` | `{ id, status, by?: human\|agent }` |
+| POST | `/api/global-todo/remove` | `{ id }` |
+| POST | `/api/global-todo/restore` | `{ todo }` — put a removed record back verbatim (undo) |
+| POST | `/api/global-todo/review` | `{ id, decision: confirm\|reject }` — answer an agent's completion |
+| GET | `/api/global-todo/report?days=7` | what the agents finished in the window, and the files it touched |
+| GET | `/api/global-todo/session?id=` | what a bound conversation last said (`tail`, `alive`, `dir`) |
+| POST | `/api/global-todo/relink` | `{ id, index, value }` — re-point a dead attachment (the value must still be a valid absolute path or URL of that kind) |
+| GET | `/api/global-todo/asset?id=&index=` | the bytes of an **image** attachment (path read from the store, never from the request; ≤8 MB, png/jpg/jpeg/webp/gif/avif) |
+| POST | `/api/global-todo/link` | `{ id, add?: [...], remove?: [value \| index] }` |
+| POST | `/api/global-todo/dispatch` | `{ id, mode?: auto\|bound\|new\|current, sessionId?, dir? }` |
+| POST | `/api/global-todo/assign` | `{ ids, items?, target?, dir? }` — hand several todos over in one call; an id that matches nothing comes back instead of being sent elsewhere |
+| POST | `/api/global-todo/settings` | `{ autoCollect?, idleMinutes?, cooldownMinutes?, exemptWorkspaces?, captureMinWeight? }` — the Host-side switches behind automatic collect; `captureMinWeight` (0–1, default 0.35) is the floor below which a batch-capture candidate is reported as skipped instead of registered |
+| POST | `/api/global-todo/diagnostics` | `{ scope, status?, reason?, message?, url? }` — the panel's own error report |
+
+`by` on `/update` and `/status` says whose write this is: omitted (what the panel sends) it is
+the person at the keyboard, and a script acting for the agent says `by: 'agent'` rather than
+being filed as a human. Anything else is a `400` — a provenance field that silently absorbs a
+typo is worse than no field at all.
+
+Errors are typed: `400` for a rejected value, `404` for an unknown id, `413` for an
+oversized body, `403` for a request that fails the fence, `502` when a hand-off cannot
+reach the agents service. The body is always `{ ok, ... }` with a machine `reason`.
+
+## Where it is listed
+
+npm is the only install source; every registry below is a discovery surface. Verified 2026-10-21 —
+re-check with `node tools/listings.mjs` rather than trusting this list.
+
+| Registry | Form | State |
+|---|---|---|
+| [npm](https://www.npmjs.com/package/dsh-global-todo) | the package | `0.11.2` |
+| [1024 Store](https://deepseek1024.com) | catalog entry ([PR #654](https://github.com/imsai-sh/awesome-deepseek-harness-plugins/pull/654), merged) | listed; its install line still resolves to the GitHub source |
+| [dsh.pub](https://dsh.pub/en/plugins/dsh-global-todo/) | published page + `dshpub` CLI | live, pinned to a commit |
+| [DSH-Plugins-Marketplace](https://github.com/bradeGithub/DSH-Plugins-Marketplace) | `registry.json`, collected from the `dsh-plugin` topic | `v0.11.2`, risk tier `safe` |
+| [awesome-dsh-plugin](https://github.com/awesome-dsh-plugin/awesome-dsh-plugin) (19.7k★) | one YAML per plugin | merged ([PR #7498](https://github.com/awesome-dsh-plugin/awesome-dsh-plugin/pull/7498)) |
+| [walkinglabs](https://github.com/walkinglabs/awesome-deepseek-harness-plugins) | source-verified README | merged ([PR #114](https://github.com/walkinglabs/awesome-deepseek-harness-plugins/pull/114)) |
+
+Five more are in review (Zhiyuan-Fan, kejixiaoliang, beancookie, dshworks, losebird) and one
+maintainer promised to include it (dshopen.org). The full table, including the auto-collected
+topic feeds we have not verified, is [docs/DISTRIBUTION.md](docs/DISTRIBUTION.md).
+
+## Where the data lives
+
+`$DSH_HOME/global-todo/todos.json`, or any absolute path you set:
+
+```yaml
+# the profile's cordis.patch.yml
+- id: global-todo
+  config:
+    storePath: /home/me/notes/global-todos.json   # a synced or versioned folder
+    dispatchDir: /home/me/projects/thing          # default cwd for a new session
+    requireExistingFiles: true                    # refuse a document that is not there yet
+    toolGroups: all                               # all (10 tools) | core (6: the list itself) | none
+```
+
+The file is the artifact: plain JSON, one record per todo, written through a temp file and
+`rename()`, with a write lock that serializes the human's and the agent's edits. A damaged
+file is reported (`store-corrupt`) rather than silently reset, and a single mangled record
+is skipped instead of taking the list down. Back it up by copying it.
+
+## Design decisions
+
+**Attachments are references, not uploads.** A document is an absolute path that stays
+where it is; an address is a URL. Nothing is copied into plugin storage, so a document
+stays editable and diffable, the list stays small, and a dispatched agent opens the very
+same path with its own tools. URLs carrying credentials are refused, because the list, the
+audit trail and every prompt would otherwise keep a copy of the secret.
+
+**Creation time is first-class.** `createdAt` is set once and never rewritten. The panel
+shows it, the tools return it as both epoch and ISO, and the dispatch prompt includes it,
+because "when did this land on my plate" is half of what a global todo is for.
+
+**One write path.** The panel and the tools both call the same service functions, so a
+validation rule, an audit entry or the write lock cannot exist on one side only.
+
+**A failed hand-off is louder than a successful one.** Success and failure are both
+recorded on the todo, and the panel shows the failure with its reason. A dispatch that
+silently did nothing is worse than one that never happened.
+
+**Optional services degrade, they do not crash.** `webServer`, `tools`, `agents`,
+`agentPresets`, `workspaceRegistry` and `agentDefaultModel` are all reached through
+`ctx.inject` or a guarded `ctx.get`. A lean composition loses exactly that surface.
+
+See [docs/DESIGN.md](docs/DESIGN.md) for the data model and the rejected alternatives,
+and [docs/ROADMAP.md](docs/ROADMAP.md) for what could come next.
+
+## Development
+
+```sh
+npm run check     # parse every file, assert the shipping shape, and — when a DSH
+                  # profile is present — validate every tool schema against the
+                  # Host's own validator
+npm test              # unit, contract and bundle tests, zero dependencies
+npm run verify        # static checks + tests
+npm run check:contracts  # the cross-field matrix a reviewer re-tests by hand:
+                      # one invalid value, one answer, whichever field it lands on
+npm run check:windows    # the same idea on a real filesystem (Windows in CI)
+npm run check:matrix  # the running dsh against engines.dsh, then the manifest,
+                      # bundle and tool-schema contracts (no preview writes)
+```
+
+`engines.dsh` is the one version fact that matters: a Host outside it skips the whole bundle
+instead of warning, so `check:matrix` is the smoke to run before widening the range.
+
+There is **no build step**. The host half is plain ESM under `lib/`; the client half is a
+classic script that registers itself with `window.__ModuleLoader__`, so no bundler,
+TypeScript toolchain or CI artifact is involved. `lib/client.js` edits take effect on a
+page refresh; host-half edits need a host restart.
+
+```sh
+npm run preview       # render the panel to preview/panel.html (+ detail and selection variants)
+npm run screenshots   # re-photograph those previews into docs/screenshots/, 1280×720
+npm run tour          # rebuild docs/screenshots/tour.gif from those PNGs (needs ffmpeg)
+```
+
+The PNGs under `docs/screenshots/` ship with the package and are embedded in both
+READMEs, so re-run `npm run screenshots` whenever the panel UI changes.
+
+```text
+lib/
+  index.js       host entry: store + service + dispatcher + the optional surfaces
+  store.js       one JSON file: atomic writes, serialized updates, tolerant loads
+  model.js       the todo entity: validation, patches, status, attachments, steps, review, notes
+  links.js       attachment rules (pure: no filesystem access)
+  service.js     the single write path, filters, sorting and the dispatch prompt
+  routes.js      the HTTP surface for the panel
+  tools.js       the ten agent tools
+  dispatch.js    handing a todo to an agent session (auto / bound / new / current)
+  completion.js  the clean-turn watcher behind automatic completion
+  capture.js     the capture duplicate gate (title, paraphrase, recent done)
+  ingest.js      the external door and its inbox rules
+  markdown.js    Markdown export and tolerant import
+  report.js      the weekly aggregate over completion records
+  reminders.js   the due rule behind the reminder strip
+  settings.js    host-side switches and the automatic-collect policy
+  metrics.js     in-memory counters reported by /health
+  diagnostics.js the bounded client error log
+  errors.js      typed failures with stable machine codes
+  client.js      the browser half (panel + sidebar entry)
+test/            unit, service, tool, route, dispatch, bundle and manifest tests
+tools/           preview.mjs — render the panel to static HTML without a Host
+                 screenshot.mjs — photograph those previews (borrows the profile's Playwright)
+                 tour.mjs — assemble the README tour GIF from those PNGs
+docs/            DESIGN.md (data model and rejected alternatives), ROADMAP.md
+```
+
+Contributions are welcome — read [CONTRIBUTING.md](CONTRIBUTING.md) first.
+
+## Feedback
+
+Questions, bug reports and ideas go to [Discussions](https://github.com/LUgABeEEE/dsh-global-todo/discussions).
+When reporting a bug, include the DSH version, the plugin version from the panel footer, your OS and
+the steps that reproduce it; a screenshot of the notice is worth more than a paragraph. Security
+issues go through [SECURITY.md](SECURITY.md) instead, not the discussion board.
+
+## Security
+
+The HTTP surface is local-only, the list is a local file, and no credential is ever stored.
+Please report vulnerabilities as described in [SECURITY.md](SECURITY.md).
+
+## License
+
+[MIT](LICENSE) © 2026 LUgABeEEE
+
+A community plugin, not a DeepSeek project: the `dsh` prefix marks the ecosystem it plugs into, not
+an endorsement. "DeepSeek" and "DeepSeek Harness" remain their owners' marks.