@lsfcy/dsh-composer-predictions 0.0.0-stage → 0.1.1

package/LICENSE

@@ -0,0 +1,21 @@
+MIT License
+
+Copyright (c) 2026 niujianya
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including without limitation the rights
+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+SOFTWARE.

package/README.en.md

@@ -0,0 +1,71 @@
+# dsh-composer-predictions
+
+[![CI](https://github.com/lsfcy/dsh-composer-predictions/actions/workflows/ci.yml/badge.svg)](https://github.com/lsfcy/dsh-composer-predictions/actions/workflows/ci.yml)
+
+Dim next-message suggestions in the DeepSeek Harness composer after a turn settles.
+**Tab** accepts the suggestion into the draft; **Esc** or typing dismisses it. Nothing is sent.
+
+English | [中文](README.md) · [Live demo](https://lsfcy.github.io/dsh-composer-predictions/demo/dsh.html?lang=en)
+
+## Example
+
+The assistant finishes with "All 12 tests pass." The empty composer then shows, dimmed:
+
+> Want me to commit this change? I can write the message.
+
+Press Tab and it becomes an editable draft. Press Esc, or start typing, and it disappears.
+
+![Composer prediction: ghost appears, Tab accepts](https://raw.githubusercontent.com/lsfcy/dsh-composer-predictions/main/demo/demo-en.gif)
+
+## Install
+
+DSH Web / desktop only (0.2.x). It does not work in other agents or harnesses.
+
+**Option 1 (recommended, inside a DSH session)**: ask the agent to run
+
+```
+plugin_manager  action: install_bundle
+                target: @lsfcy/dsh-composer-predictions
+```
+
+It applies without a restart (refresh the page if the UI does not update).
+
+**Option 2 (terminal)**:
+
+```bash
+dsh plugin --profile desktop add @lsfcy/dsh-composer-predictions
+```
+
+Restart DSH afterwards. The desktop app uses `desktop`; other entrypoints use their profile under `~/.dsh/profiles`, e.g. `web`. To pin a version without npm:
+`dsh plugin --profile desktop add git+https://github.com/lsfcy/dsh-composer-predictions.git#v0.1.1`.
+
+## Config
+
+**Settings → 输入框预测 / Composer prediction**:
+
+| Key | Default | Meaning |
+| --- | --- | --- |
+| `enabled` | `true` | Master switch: off stops requests and hides the ghost. |
+| `provider` | auto | Empty uses the provider you are currently using. |
+| `model` | auto | Empty uses the cheapest model on that provider. |
+| `reasoningEffort` | auto | Empty uses the cheapest effort the model supports. |
+
+The same keys work as the bundle row's `config` in a profile patch.
+
+## How it works
+
+1. The host reads the session's real user/assistant messages, keeps the last 12, and clips them to 6,000 characters.
+2. It only predicts after a completed turn that ended on an assistant reply.
+3. One cheap model call returns a single line (≤160 characters), cached per session content for 30 minutes.
+4. The client paints it as ghost text; Tab writes it into the draft.
+
+## Cost
+
+One generation uses roughly 2–3k tokens (system prompt + transcript + output capped at 512) on the cheapest model at the lowest reasoning effort. Cache hits are free, and aborted turns never call the model.
+
+## Limits
+
+- It is a guess, not memory: no retrieval, no long-term profile, no knowledge beyond the recent transcript.
+- Only after successful turns; a suggestion never sends itself.
+- Cheap models can be bland or wrong. Pin a stronger model in Config if you care more about quality.
+- One line only, no multi-step plans.

package/README.md

@@ -1,3 +1,72 @@
-# Temporary Holding Version
+# dsh-composer-predictions
 
-This version is a temporary placeholder for this package. An operational version to replace this has been submitted for review and is awaiting a staged release.
+[![CI](https://github.com/lsfcy/dsh-composer-predictions/actions/workflows/ci.yml/badge.svg)](https://github.com/lsfcy/dsh-composer-predictions/actions/workflows/ci.yml)
+
+给 DeepSeek Harness 的输入框补上“下一句预测”：一轮对话结束后，在空的输入框里用灰字显示
+你接下来可能发送的消息。**Tab** 填入草稿，**Esc** 或开始打字即忽略。不会自动发送。
+
+[English](README.en.md) | 中文 · [在线试玩](https://lsfcy.github.io/dsh-composer-predictions/demo/dsh.html)
+
+## 示例
+
+助手刚说完「测试全部通过」，空输入框里出现灰字：
+
+> 顺手把这次改动提交了吗？提交信息我一起写好
+
+按 Tab，它变成可编辑的草稿；按 Esc 或直接打字，它消失。
+
+![灰字预测：出现、Tab 接受](https://raw.githubusercontent.com/lsfcy/dsh-composer-predictions/main/demo/demo-zh.gif)
+
+## 安装
+
+仅适用于 DSH 的 Web / 桌面端（0.2.x），不能用于其他 agent/harness。
+
+**方式一（推荐，已在 DSH 会话里）**：让 agent 执行
+
+```
+plugin_manager  action: install_bundle
+                target: @lsfcy/dsh-composer-predictions
+```
+
+即时生效，不用重启（页面没更新就刷新一次）。
+
+**方式二（终端）**：
+
+```bash
+dsh plugin --profile desktop add @lsfcy/dsh-composer-predictions
+```
+
+装完重启 DSH。桌面端 profile 是 `desktop`；其他入口用 `~/.dsh/profiles` 里对应的名字，如 `web`。不想走 npm 时，也可以用固定版本：
+`dsh plugin --profile desktop add git+https://github.com/lsfcy/dsh-composer-predictions.git#v0.1.1`。
+
+## 配置
+
+**设置 → 输入框预测**：
+
+| 配置 | 默认 | 说明 |
+| --- | --- | --- |
+| `enabled` | `true` | 总开关：关闭后不再请求，也不显示灰字。 |
+| `provider` | 自动 | 留空用你当前选中的 provider。 |
+| `model` | 自动 | 留空用该 provider 下最便宜的型号。 |
+| `reasoningEffort` | 自动 | 留空用模型支持的最省推理档。 |
+
+这些键同样可以写在 profile patch 的 bundle row `config` 里。
+
+## 原理
+
+1. 宿主读取会话里真人的 user/assistant 消息，只取最近 12 条，压到 6000 字符。
+2. 只在一轮**成功结束**、且最后一条是 assistant 回复时预测。
+3. 调一次便宜模型，取一行（≤160 字符），按会话内容缓存 30 分钟。
+4. 客户端把它画成灰字；Tab 写入输入框。
+
+## 成本
+
+单次生成约 2–3k token（system 提示词 + 转录 + 输出上限 512），走最便宜型号和最低推理档。
+命中缓存不产生费用；被中止的轮次不会调用模型。
+
+## 限制
+
+- 它只是猜测，没有记忆：不做检索，不读长期画像，只看最近这段转录。
+- 只在成功轮次后出现，且永不自动发送。
+- 便宜模型可能平淡或猜错；在意质量可在配置里固定更强的模型。
+- 只给一行，不会给多步计划。

package/build.mjs

@@ -0,0 +1,26 @@
+// Concatenate the two browser sources into the lazy-CJS bundle protocol the
+// DSH client module loader expects (window.__ModuleLoader__.load). No bundler,
+// no dependencies: src/ghost.js and src/main.js are plain scripts that run
+// inside the factory scope, so `require` there resolves against the platform
+// module table (react, react/jsx-runtime, ...).
+import { readFileSync, writeFileSync } from 'node:fs'
+
+const packageJson = JSON.parse(readFileSync(new URL('./package.json', import.meta.url), 'utf8'))
+const ghost = readFileSync(new URL('./src/ghost.js', import.meta.url), 'utf8')
+const main = readFileSync(new URL('./src/main.js', import.meta.url), 'utf8')
+
+const bundle = `// GENERATED FILE — edit src/ghost.js or src/main.js and run: node build.mjs
+window.__ModuleLoader__.load({
+  id: ${JSON.stringify(packageJson.name)},
+  factory: (require) => {
+    var module = { exports: {} }
+    var exports = module.exports
+${ghost}
+${main}
+    return module.exports
+  },
+})
+`
+
+writeFileSync(new URL('./client.js', import.meta.url), bundle)
+console.log('client.js written (' + bundle.length + ' bytes)')