freecodego 0.1.6-alpha.2.4 → 0.1.7-alpha.2.1

package/README.i18n.yaml

@@ -2,5 +2,5 @@
 # side as of the last confirmed-consistent state. Both languages carry equal authority;
 # after editing either side, bring the other along and re-record with:
 #   pnpm run verify-translation-pairing --write packages/freecodego/bundle-latest/README.md
-README.md: a55282f22dea6b39738bf359c0aa3942ae935bc9
-README.zh.md: 6892acf2dc9a8ef0ccdc5f88b482d335f113320a
+README.md: 0b00540c87ee0781c697d407579b7c2617190d21
+README.zh.md: 9be8ba5e80b702e0f0b442933b0e35a9643e073d

package/README.md

@@ -9,7 +9,7 @@ English | [中文](README.zh.md)
 
 ## Summary
 
-Installable single-package FreeCodeGo composition for DeepSeek Harness `freecodego@0.1.6-alpha.2`, targeting Harness baseline `0.1.6-alpha.2`. The npm artifact contains the compiled Host plugin, browser client, session-event prerequisite, native worker entrypoints, and the bundled Harness Agent Teams composition. A release publishes on npm's `next` dist-tag, and its plugin version intentionally matches the Harness version it targets exactly; for a reproducible install, name the version (`freecodego@0.1.6-alpha.2`) instead of the channel. Official Codex and Claude runtime binaries remain optional platform downloads — the package does not embed every platform's native binary.
+Installable single-package FreeCodeGo composition for DeepSeek Harness `freecodego@0.1.7-alpha.2.1`, targeting Harness baseline `0.1.7-alpha.2`. The npm artifact contains the compiled Host plugin, browser client, session-event prerequisite, native worker entrypoints, and the bundled Harness Agent Teams composition. A release publishes on npm's `next` dist-tag, and its plugin version intentionally matches the Harness version it targets exactly; for a reproducible install, name the version (`freecodego@0.1.7-alpha.2`) instead of the channel. Official Codex and Claude runtime binaries remain optional platform downloads — the package does not embed every platform's native binary.
 
 ## Table of Contents
 
@@ -18,6 +18,7 @@ Installable single-package FreeCodeGo composition for DeepSeek Harness `freecode
 - [Code Review](#code-review)
 - [Project Memory](#project-memory)
 - [Agent Teams](#agent-teams)
+- [Image and Video Generation](#image-and-video-generation)
 - [Model Experience](#model-experience)
 - [Known Limitations and Deferred Work](#known-limitations-and-deferred-work)
 - [Dev Note](#dev-note)
@@ -31,10 +32,10 @@ Install the release for the Harness baseline you run; each release carries one t
 
 ```sh
 dsh plugin --profile web add --save-exact \
-  https://github.com/XiangSu-ce/dsh-plugin-freecodego/releases/download/freecodego-v0.1.6-alpha.2/freecodego-0.1.6-alpha.2.tgz
+  https://github.com/XiangSu-ce/dsh-plugin-freecodego/releases/download/freecodego-v0.1.7-alpha.2.1/freecodego-0.1.7-alpha.2.tgz
 ```
 
-`freecodego-0.1.6-alpha.2.tgz` is the bundle for Harness `0.1.6-alpha.2`: substitute the version you run, which the settings page shows beside the installed plugin version. The asset name is the naming contract in `packages/freecodego/AGENTS.md`.
+`freecodego-0.1.7-alpha.2.tgz` is the bundle for Harness `0.1.7-alpha.2`: substitute the version you run, which the settings page shows beside the installed plugin version. The asset name is the naming contract in `packages/freecodego/AGENTS.md`.
 
 The `dsh` command is provided by the Harness CLI, not this bundle. In a normal terminal install it first with `npm install --global @deepseek-ai/dsh` (and ensure `pnpm` is available). Desktop launches the same command through its private shims and passes the active `DSH_HOME`, so Web and Desktop use one Profile data directory when they select the same Harness home.
 
@@ -57,9 +58,9 @@ Every free row below comes from the provider's own directory, read when you open
 
 | Provider | Free models | Directory |
 |---|---|---|
-| **OpenCode** | `big-pickle`, `deepseek-v4-flash-free`, `jev-1.13-free`, `ling-3.0-flash-fin-free`, `mimo-v2.5-free`, `mimo-v2.6-flash-free`, `muse-spark-1.2`, `muse-spark-1.2-contributor-free`, `muse-spark-1.3`, `muse-spark-1.3-contributor-free`, `nemotron-3-ultra-free`, `nemotron-3.5-lightning-free` | 12 of 76 rows; public, no sign-in |
-| **Kilo** | `cohere/north-mini-code:free`, `dots-studio/dots-3-note-preview:free`, `inclusionai/ling-3.0-flash-fin:free`, `inclusionai/ling-3.0-flash-sante:free`, `inclusionai/ling-3.0-flash-vl:free`, `kilo-auto/free`, `liquid/lfm-2.5-2.6b:free`, `nex-agi/nex-n2.5-mini:free`, `nex-agi/nex-n2.5-pro:free`, `nvidia/nemotron-3-nano-omni-30b-a3b-reasoning:free`, `nvidia/nemotron-3-super-120b-a12b:free`, `nvidia/nemotron-3-ultra-550b-a55b:free`, `nvidia/nemotron-3.5-content-safety:free`, `nvidia/nemotron-3.5-lightning:free`, `openrouter/free`, `poolside/laguna-s-2.1:free`, `poolside/laguna-xs-2.1:free`, `qwen/qwen3.8-27b:free`, `stepfun/step-3.7-flash:free`, `thinkingmachines/inkling-small:free`, `z-ai/glm-5.2:free` | 21 of 385 rows; public, 200 requests/hour per egress IP |
-| **Logfare** | chat `deepseek-v3.2`, `deepseek-v4-pro-0813`, `gemma-4-26b`, `gemma-4-31b-it`, `glm-5`, `glm-5.3`, `glm-5.3-flash`, `grok-4.6`, `kimi-k2.5`, `kimi-k2.6`, `kimi-k2.7-code`, `logfare/auto`, `moondream3.1`, `qwen-3.8-27b`, `step-3.7-flash`; images `flux-1-schnell`, `flux-2-dev`, `flux-2-klein-4b`, `flux-2-klein-9b`, `sdxl-lightning`; audio `melotts`, `whisper-large-v3-turbo`; other routes `aura-2-en`, `lucid-origin`, `nova-3`, `phoenix-1.0` | 26 rows; 18 need a training-data opt-in, the other 8 do not |
+| **OpenCode** | `big-pickle`, `deepseek-v4-flash-free`, `jev-1.13-free`, `ling-3.0-flash-fin-free`, `mimo-v2.5-free`, `mimo-v2.6-flash-free`, `muse-spark-1.2`, `muse-spark-1.2-contributor-free`, `muse-spark-1.3`, `muse-spark-1.3-contributor-free`, `nemotron-3-ultra-free`, `nemotron-3.5-lightning-free`, `space-bunny-free` | 13 of 80 rows; public, no sign-in |
+| **Kilo** | `cohere/north-mini-code:free`, `dots-studio/dots-3-note-preview:free`, `inclusionai/ling-3.0-flash-fin:free`, `inclusionai/ling-3.0-flash-sante:free`, `inclusionai/ling-3.0-flash-vl:free`, `kilo-auto/free`, `liquid/lfm-2.5-2.6b:free`, `nex-agi/nex-n2.5-mini:free`, `nex-agi/nex-n2.5-pro:free`, `nvidia/nemotron-3-nano-omni-30b-a3b-reasoning:free`, `nvidia/nemotron-3-super-120b-a12b:free`, `nvidia/nemotron-3-ultra-550b-a55b:free`, `nvidia/nemotron-3.5-content-safety:free`, `nvidia/nemotron-3.5-lightning:free`, `openrouter/free`, `poolside/laguna-s-2.1:free`, `poolside/laguna-xs-2.1:free`, `qwen/qwen3.8-27b:free`, `stepfun/step-3.7-flash:free`, `thinkingmachines/inkling-small:free`, `z-ai/glm-5.2:free` | 21 of 394 rows; public, 200 requests/hour per egress IP |
+| **Logfare** | chat `claude-opus-4.6`, `deepseek-v3.2`, `deepseek-v4-pro-0813`, `gemma-4-26b`, `glm-5`, `glm-5.3`, `glm-5.3-flash`, `grok-4.6`, `kimi-k2.5`, `kimi-k2.6`, `kimi-k2.7-code`, `kimi-k3`, `logfare/auto`, `moondream3.1`, `qwen-3.8-27b`, `step-3.7-flash`; images `flux-1-schnell`, `flux-2-dev`, `flux-2-klein-4b`, `flux-2-klein-9b`, `sdxl-lightning`; audio `melotts`, `whisper-large-v3-turbo`; other routes `aura-2-en`, `lucid-origin`, `nova-3`, `phoenix-1.0` | 27 rows; 20 need a training-data opt-in, the other 7 do not |
 | **Qoder** | `Qwen 3.8 Flash` (route `qmodel_38flash`) | the free flash route, plus daily check-in campaigns |
 | **NVIDIA** | `google/gemma-4-31b-it`, `moonshotai/kimi-k3`, `z-ai/glm-5.3`, `z-ai/glm-5.3-flash` — the roster also names `deepseek-ai/deepseek-v4-flash-0731` and `deepseek-ai/deepseek-v4-pro-0813`, which are gone from NVIDIA's live catalogue of 82 rows | an API key is required to call them; cross-checked 2026-09-23 |
 | **SenseNova** | `deepseek-v4-flash`, `deepseek-v4-pro`, `glm-5.2`, `kimi-k3`, `sensenova-6.8-flash-lite` — 1M context and 128K output each | roster ships in the bundle; an API key is required |
@@ -74,7 +75,7 @@ These lists follow their directories: a route upstream retires leaves the table
 
 TRAE, Cline, WorkBuddy International, Agnes publish no stable roster, so their rows are counted when they arrive rather than listed here.
 
-18 Logfare rows sit behind a training-data opt-in, which the picker labels rather than hides.
+20 Logfare rows sit behind a training-data opt-in, which the picker labels rather than hides.
 
 <!-- generated:free-models:end -->
 
@@ -95,6 +96,15 @@ Durable per-project memory ships with the bundle: recall is reviewed-only, injec
 
 The bundle enables Harness Agent Teams with same-engine child sessions. A DeepSeek parent creates DeepSeek teammates, a Codex parent creates Codex teammates, and a Claude parent creates Claude teammates; the child route is inherited from the parent session rather than selected from the process default. Team roster, mailbox, task-board, and teammate Session records remain owned by Harness. The Web Chat renders the live delegated-Agent progress tree inside the parent conversation, including each task label, current tool, status, and tool-use count.
 
+<a id="image-and-video-generation"></a>
+## Image and Video Generation
+
+The bundle registers image and video generation tools for DeepSeek, Codex, and Claude, and they read the model each category is configured to use at execution time rather than a pinned one. `mediaGenerationEnabled` (on by default) is the switch: with it off, `freecodego_generate_image`, `freecodego_generate_video`, and the legacy `agnes_generate_image` / `agnes_generate_video` names are **unregistered** — they leave the model's tool list instead of failing at call time — while `freecodego_generate_audio` and `freecodego_transcribe_audio` stay, since one writes a file into the active workspace and the other reads one out of it.
+
+The route is chosen from the **provider**, never from a keyword in a model id. Image generation speaks the OpenAI Images contract (`/images/generations`, and `/images/edits` when source images are attached), Seedream's `image: [...]` fusion body, Gemini's `:generateContent`, Imagen's `:predict`, and DashScope's `multimodal-generation`. Video generation speaks nine: Kling's `text2video` / `image2video` / `multi-image2video`, Ark's `/contents/generations/tasks`, DashScope's asynchronous `video-synthesis`, MiniMax Hailuo's `/v2/video_generation`, Vidu's four `ent/v2` routes, Gemini's Veo `:predictLongRunning`, xAI's `/videos/generations` (with edits and extensions when a source video is given), OpenAI's `/videos`, and the gateway's portable shape for everything unrecognized.
+
+The model roster is **live, not pinned**: candidates are collected per request from the Harness model directory, the managed catalog, Logfare, and Agnes, and a route's category is inferred from its id and can be overridden by hand in the settings page's model categories. Every call carries a per-route circuit breaker, and only a route's own problem falls through to the next one — a wrong credential or a refused prompt fails outright rather than paying for another attempt. The package README in `harness-plugin` carries the full table, including which providers accept a last frame and which durations each route can render.
+
 -----
 
 <a id="model-experience"></a>

package/README.zh.md

@@ -9,7 +9,7 @@ kind: "package-bundle"
 
 ## 概述
 
-面向 DeepSeek Harness `freecodego@0.1.6-alpha.2` 的可安装单包 FreeCodeGo 组合，目标 Harness 基线为 `0.1.6-alpha.2`。npm 产物包含编译后的 Host 插件、浏览器端 client、会话事件前置包、原生 worker 入口，以及随包交付的 Harness Agent Teams 组合。发布产物使用 npm 的 `next` dist-tag，且其插件版本刻意与所面向的 Harness 版本完全一致；需要可复现安装时请指定版本（`freecodego@0.1.6-alpha.2`），而不是渠道名。官方 Codex 与 Claude 运行时二进制仍按平台可选下载——本包不内嵌每个平台的原生二进制。
+面向 DeepSeek Harness `freecodego@0.1.7-alpha.2.1` 的可安装单包 FreeCodeGo 组合，目标 Harness 基线为 `0.1.7-alpha.2`。npm 产物包含编译后的 Host 插件、浏览器端 client、会话事件前置包、原生 worker 入口，以及随包交付的 Harness Agent Teams 组合。发布产物使用 npm 的 `next` dist-tag，且其插件版本刻意与所面向的 Harness 版本完全一致；需要可复现安装时请指定版本（`freecodego@0.1.7-alpha.2`），而不是渠道名。官方 Codex 与 Claude 运行时二进制仍按平台可选下载——本包不内嵌每个平台的原生二进制。
 
 ## 目录
 
@@ -18,6 +18,7 @@ kind: "package-bundle"
 - [代码审查](#code-review)
 - [项目记忆](#project-memory)
 - [Agent Teams](#agent-teams)
+- [生图与视频生成](#image-and-video-generation)
 - [模型体验](#model-experience)
 - [已知限制与延期工作](#known-limitations-and-deferred-work)
 - [开发备注](#dev-note)
@@ -31,10 +32,10 @@ kind: "package-bundle"
 
 ```sh
 dsh plugin --profile web add --save-exact \
-  https://github.com/XiangSu-ce/dsh-plugin-freecodego/releases/download/freecodego-v0.1.6-alpha.2/freecodego-0.1.6-alpha.2.tgz
+  https://github.com/XiangSu-ce/dsh-plugin-freecodego/releases/download/freecodego-v0.1.7-alpha.2.1/freecodego-0.1.7-alpha.2.tgz
 ```
 
-`freecodego-0.1.6-alpha.2.tgz` 就是面向 Harness `0.1.6-alpha.2` 的 bundle：请替换为你实际运行的版本，设置页会把它显示在已装插件版本旁边。资产名即 `packages/freecodego/AGENTS.md` 里的命名约定。
+`freecodego-0.1.7-alpha.2.tgz` 就是面向 Harness `0.1.7-alpha.2` 的 bundle：请替换为你实际运行的版本，设置页会把它显示在已装插件版本旁边。资产名即 `packages/freecodego/AGENTS.md` 里的命名约定。
 
 `dsh` 命令由 Harness CLI 提供，不来自本 bundle。普通终端里请先用 `npm install --global @deepseek-ai/dsh` 安装它（并确保 `pnpm` 可用）。桌面端通过自己的私有 shim 运行同一条命令并传入当前 `DSH_HOME`，因此 Web 与桌面端选择同一个 Harness home 时共用同一份 Profile 数据目录。
 
@@ -57,9 +58,9 @@ gh workflow run release-freecodego.yml --ref freecodego-v<version>
 
 | 提供商 | 免费模型 | 目录 |
 |---|---|---|
-| **OpenCode** | `big-pickle`、`deepseek-v4-flash-free`、`jev-1.13-free`、`ling-3.0-flash-fin-free`、`mimo-v2.5-free`、`mimo-v2.6-flash-free`、`muse-spark-1.2`、`muse-spark-1.2-contributor-free`、`muse-spark-1.3`、`muse-spark-1.3-contributor-free`、`nemotron-3-ultra-free`、`nemotron-3.5-lightning-free` | 76 行中的 12 行；公开，无需登录 |
-| **Kilo** | `cohere/north-mini-code:free`、`dots-studio/dots-3-note-preview:free`、`inclusionai/ling-3.0-flash-fin:free`、`inclusionai/ling-3.0-flash-sante:free`、`inclusionai/ling-3.0-flash-vl:free`、`kilo-auto/free`、`liquid/lfm-2.5-2.6b:free`、`nex-agi/nex-n2.5-mini:free`、`nex-agi/nex-n2.5-pro:free`、`nvidia/nemotron-3-nano-omni-30b-a3b-reasoning:free`、`nvidia/nemotron-3-super-120b-a12b:free`、`nvidia/nemotron-3-ultra-550b-a55b:free`、`nvidia/nemotron-3.5-content-safety:free`、`nvidia/nemotron-3.5-lightning:free`、`openrouter/free`、`poolside/laguna-s-2.1:free`、`poolside/laguna-xs-2.1:free`、`qwen/qwen3.8-27b:free`、`stepfun/step-3.7-flash:free`、`thinkingmachines/inkling-small:free`、`z-ai/glm-5.2:free` | 385 行中的 21 行；公开，每个出口 IP 每小时 200 次 |
-| **Logfare** | 对话 `deepseek-v3.2`、`deepseek-v4-pro-0813`、`gemma-4-26b`、`gemma-4-31b-it`、`glm-5`、`glm-5.3`、`glm-5.3-flash`、`grok-4.6`、`kimi-k2.5`、`kimi-k2.6`、`kimi-k2.7-code`、`logfare/auto`、`moondream3.1`、`qwen-3.8-27b`、`step-3.7-flash`；图片 `flux-1-schnell`、`flux-2-dev`、`flux-2-klein-4b`、`flux-2-klein-9b`、`sdxl-lightning`；音频 `melotts`、`whisper-large-v3-turbo`；其他路由 `aura-2-en`、`lucid-origin`、`nova-3`、`phoenix-1.0` | 26 行；18 行需要训练数据授权，其余 8 行不需要 |
+| **OpenCode** | `big-pickle`、`deepseek-v4-flash-free`、`jev-1.13-free`、`ling-3.0-flash-fin-free`、`mimo-v2.5-free`、`mimo-v2.6-flash-free`、`muse-spark-1.2`、`muse-spark-1.2-contributor-free`、`muse-spark-1.3`、`muse-spark-1.3-contributor-free`、`nemotron-3-ultra-free`、`nemotron-3.5-lightning-free`、`space-bunny-free` | 80 行中的 13 行；公开，无需登录 |
+| **Kilo** | `cohere/north-mini-code:free`、`dots-studio/dots-3-note-preview:free`、`inclusionai/ling-3.0-flash-fin:free`、`inclusionai/ling-3.0-flash-sante:free`、`inclusionai/ling-3.0-flash-vl:free`、`kilo-auto/free`、`liquid/lfm-2.5-2.6b:free`、`nex-agi/nex-n2.5-mini:free`、`nex-agi/nex-n2.5-pro:free`、`nvidia/nemotron-3-nano-omni-30b-a3b-reasoning:free`、`nvidia/nemotron-3-super-120b-a12b:free`、`nvidia/nemotron-3-ultra-550b-a55b:free`、`nvidia/nemotron-3.5-content-safety:free`、`nvidia/nemotron-3.5-lightning:free`、`openrouter/free`、`poolside/laguna-s-2.1:free`、`poolside/laguna-xs-2.1:free`、`qwen/qwen3.8-27b:free`、`stepfun/step-3.7-flash:free`、`thinkingmachines/inkling-small:free`、`z-ai/glm-5.2:free` | 394 行中的 21 行；公开，每个出口 IP 每小时 200 次 |
+| **Logfare** | 对话 `claude-opus-4.6`、`deepseek-v3.2`、`deepseek-v4-pro-0813`、`gemma-4-26b`、`glm-5`、`glm-5.3`、`glm-5.3-flash`、`grok-4.6`、`kimi-k2.5`、`kimi-k2.6`、`kimi-k2.7-code`、`kimi-k3`、`logfare/auto`、`moondream3.1`、`qwen-3.8-27b`、`step-3.7-flash`；图片 `flux-1-schnell`、`flux-2-dev`、`flux-2-klein-4b`、`flux-2-klein-9b`、`sdxl-lightning`；音频 `melotts`、`whisper-large-v3-turbo`；其他路由 `aura-2-en`、`lucid-origin`、`nova-3`、`phoenix-1.0` | 27 行；20 行需要训练数据授权，其余 7 行不需要 |
 | **Qoder** | `Qwen 3.8 Flash`（路由 `qmodel_38flash`） | 免费 flash 路由，另有每日签到活动 |
 | **NVIDIA** | `google/gemma-4-31b-it`、`moonshotai/kimi-k3`、`z-ai/glm-5.3`、`z-ai/glm-5.3-flash` —— 名单里另有 `deepseek-ai/deepseek-v4-flash-0731` 与 `deepseek-ai/deepseek-v4-pro-0813`，这两个名字已不在 NVIDIA 实时目录（82 行）中 | 调用需要 API key；核对于 2026-09-23 |
 | **SenseNova** | `deepseek-v4-flash`、`deepseek-v4-pro`、`glm-5.2`、`kimi-k3`、`sensenova-6.8-flash-lite` —— 均为 1M 上下文 / 128K 输出 | 名单随 bundle 内置；需要 API key |
@@ -74,7 +75,7 @@ gh workflow run release-freecodego.yml --ref freecodego-v<version>
 
 TRAE、Cline、WorkBuddy 国际版、Agnes 不公布固定名单，因此它们的行在到达时计数，而不在此列名。
 
-Logfare 有 18 行位于训练数据授权之后，选择器会标注而不是隐藏它们。
+Logfare 有 20 行位于训练数据授权之后，选择器会标注而不是隐藏它们。
 
 <!-- generated:free-models:end -->
 
@@ -95,6 +96,15 @@ Logfare 有 18 行位于训练数据授权之后，选择器会标注而不是
 
 bundle 以同引擎子会话启用 Harness Agent Teams。DeepSeek 父会话创建 DeepSeek 队友，Codex 父会话创建 Codex 队友，Claude 父会话创建 Claude 队友；子会话路由继承自父会话，而不是从进程默认值中选取。团队名册、邮箱、任务看板与队友 Session 记录仍由 Harness 持有。Web Chat 在父对话内渲染实时的委派 Agent 进度树，包括每个任务标签、当前工具、状态与工具使用次数。
 
+<a id="image-and-video-generation"></a>
+## 生图与视频生成
+
+bundle 为 DeepSeek、Codex 与 Claude 注册生图与生视频工具，执行时读取这一类目配置的模型，而不是钉死某个模型。`mediaGenerationEnabled`（默认开启）就是那个开关：关掉后 `freecodego_generate_image`、`freecodego_generate_video` 与旧名字 `agnes_generate_image` / `agnes_generate_video` 会被**注销**——它们从模型的工具表里整体离开，而不是等到调用时才报错——而 `freecodego_generate_audio` 与 `freecodego_transcribe_audio` 保留，因为一个往活动工作区写文件、一个从活动工作区读文件。
+
+路线由 **provider** 决定，绝不由模型 id 里的关键词决定。生图讲 OpenAI Images 契约（`/images/generations`，带源图时走 `/images/edits`）、Seedream 的 `image: [...]` 融合体、Gemini 的 `:generateContent`、Imagen 的 `:predict`，以及 DashScope 的 `multimodal-generation`。生视频讲九条：Kling 的 `text2video` / `image2video` / `multi-image2video`、Ark 的 `/contents/generations/tasks`、DashScope 的异步 `video-synthesis`、MiniMax Hailuo 的 `/v2/video_generation`、Vidu 的四条 `ent/v2` 路由、Gemini 的 Veo `:predictLongRunning`、xAI 的 `/videos/generations`（给源视频时含编辑与续写）、OpenAI 的 `/videos`，以及网关对一切未识别 provider 的通用形状。
+
+模型名单是**活的，不是钉死的**：候选每次请求实时从 Harness 模型目录、托管目录、Logfare 与 Agnes 汇总，路线分类按 id 推断，也可在设置页的模型分类里手动覆盖。每次调用都带逐路线熔断，只有路线自己的问题才会落到下一条——密钥错或 prompt 被拒直接失败，而不是再付一次尝试的钱。完整表格（包括哪些 provider 接受尾帧、每条路线能渲染多长）见 `harness-plugin` 的包内 README。
+
 -----
 
 <a id="model-experience"></a>

package/cordis.patch.yml

@@ -1,4 +1,4 @@
-# Standard DeepSeek Harness 0.1.6-alpha.2 FreeCodeGo composition.
+# Standard DeepSeek Harness 0.1.7-alpha.2 FreeCodeGo composition.
 # Neither official route is disabled. `llm-deepseek` is the only row that
 # registers provider `deepseek-official`, and the base composition names that
 # provider twice (`agent-default-model.provider`, `web.searchProvider`) — so
@@ -96,11 +96,15 @@
     #    fails on an install that lacks it, so these rows are `disabled` and carry
     #    the recipe for whoever wants them.
     # 2. **They fail closed at different grains, so the grain decides.** Mounted
-    #    in a *preset*, one unresolvable enabled row makes the whole preset
-    #    unselectable — `packages/preset/agent-presets/src/discovery.ts`,
-    #    `unresolvableRows`, skips a row only while `Boolean(row.disabled)` — which
-    #    is the "Failed to load" failure the bundled presets already record once.
-    #    Mounted here, an enabled row that cannot load fails *that entry alone*:
+    #    in a *preset*, one unusable enabled row makes the whole preset
+    #    unselectable: 0.1.6 reached that verdict in
+    #    `packages/preset/agent-presets/src/discovery.ts` (`unresolvableRows`,
+    #    which skipped a row only while `Boolean(row.disabled)`), and 0.1.7
+    #    reaches the same one in `packages/preset/agent-preset-registry/src/mount.ts`
+    #    (`auditRows`, which reports import failures, activation failures, and
+    #    rows still waiting for a service). Either way it is the "Failed to
+    #    load" failure the bundled presets already record once. Mounted here, an
+    #    enabled row that cannot load fails *that entry alone*:
     #    `Entry._init` catches the import error, logs it and returns, and the rest
     #    of the tree keeps running. The host plane is therefore the plane that can
     #    hold an optional capability at all.
@@ -186,6 +190,38 @@
       name: '@deepseek-ai/dsh-tool-session-query'
       disabled: true
 
+    # ── the Harness plugin's own row ───────────────────────────────────────
+    #
+    # The plugin is a Loader entry, not a programmatic mount from the `freecodego`
+    # row below. Being an entry is what makes it configurable, and that is the
+    # whole difference: the settings service addresses a plugin by its profile
+    # entry id, reads the live document from `entry.fiber.config`, and validates a
+    # write against that entry's own `Config`.
+    #
+    # `SettingsForms.describe`/`write` resolve an entry through
+    # `configEditor.entries()` and drop any entry whose fiber has no runtime
+    # `Config` (`packages/settings/settings/src/index.ts`). Mounted with
+    # `ctx.plugin()` inside the `freecodego` row's `apply`, this plugin inherited
+    # the *bundle's* entry: `ctx.fiber.entry.options.id` was `freecodego`, that
+    # entry has no schema, and every write was refused with `No configurable
+    # plugin entry "freecodego"` while every read answered the empty document.
+    # Nothing in this plugin could be turned on from the FreeCodeGo settings tab,
+    # and nothing stayed on across a restart.
+    #
+    # Three readers depend on this id, and they are the ones to check before
+    # renaming the row: `harness-ui` reads and writes it (its
+    # `tests/settings-entry.spec.ts` pins the client constant to this line),
+    # `plugin-conflicts.ts` reads the switch out of this row's `config` before the
+    # Loader starts, and the settings service appends its own override row here.
+    #
+    # Position is last in this layer on purpose, which is exactly where the
+    # programmatic mount used to happen: the rows above are mounted *before* it,
+    # and the team runtime, the browser/desktop providers, and the session-history
+    # tools each document that position as a requirement of their own.
+    - id: freecodego-harness-plugin
+      name: 'freecodego/harness-plugin'
+      config: {}
+
     - id: freecodego
       name: 'freecodego'
       config: {}

package/dist/agent-team.js

@@ -14875,7 +14875,7 @@ var TeamRoster = class {
       id: root.id,
       name: "lead",
       role: "lead",
-      status: root.status,
+      status: availability(root),
       ...root.options.model === void 0 ? {} : { model: root.options.model },
       diagnostics: []
     }];
@@ -14886,7 +14886,7 @@ var TeamRoster = class {
         id: member.id,
         name: member.name,
         role: "teammate",
-        status: member.phase === "failed" ? "failed" : member.phase === "provisioning" ? "provisioning" : live?.status ?? "inactive",
+        status: member.phase === "failed" ? "failed" : member.phase === "provisioning" ? "provisioning" : availability(live),
         description: member.description,
         provider: member.provider,
         context: member.context,
@@ -14943,7 +14943,7 @@ var TeamRoster = class {
     if (target.id === membership.root.id) throw new TeamError("the Team Lead cannot interrupt itself", "TEAM_INVALID_TARGET");
     const live = this.ctx.agents.get(target.id);
     if (live === void 0) return { previousStatus: "inactive" };
-    const previousStatus = live.status;
+    const previousStatus = availability(live);
     this.ctx.subagents.interrupt(target.id, {
       kind: "ancestor",
       agent: caller
@@ -15125,7 +15125,7 @@ var TeamRoster = class {
       id: member.id,
       name: member.name,
       role: "teammate",
-      status: live?.status ?? "inactive",
+      status: availability(live),
       description: member.description,
       provider: member.provider,
       context: member.context,
@@ -15157,6 +15157,9 @@ var TeamRoster = class {
     return foldSubagentDescriptor(agent.session.snapshotEvents(agent.session.inheritedEventCount)) !== void 0;
   }
 };
+function availability(agent) {
+  return agent?.status === "running" ? "running" : "inactive";
+}
 var TeamMailbox = class {
   ctx;
   journal;
@@ -15497,13 +15500,11 @@ var contentBlockSchema = external_exports.lazy(() => external_exports.union([
     name: external_exports.string(),
     arguments: external_exports.string()
   }).strict(),
-  external_exports.object({
-    type: external_exports.literal("tool-result"),
-    toolCallId: external_exports.string().min(1),
-    content: external_exports.array(contentBlockSchema),
-    isError: external_exports.boolean().optional()
-  }).strict(),
-  external_exports.object({ type: external_exports.string().min(1) }).loose().refine((block) => !coreContentBlockTypes.has(block.type), { message: "known content block types must match their declared fields" })
+  external_exports.custom((value) => {
+    if (typeof value !== "object" || value === null || Array.isArray(value)) return false;
+    const type = value.type;
+    return typeof type === "string" && type.length > 0 && !coreContentBlockTypes.has(type);
+  })
 ]));
 var teamMemberSnapshotSchema = external_exports.object({
   id: sessionIdSchema,
@@ -15688,7 +15689,7 @@ function applyCurrentTeamEvent(state, event) {
 }
 var teamProjectionDefinition = {
   key: "agentTeam",
-  stateVersion: 3,
+  stateVersion: 4,
   stateSchema: teamProjectionEntrySchema,
   init: (header) => emptyTeamState(header.id),
   apply: (state, event) => {
@@ -16002,14 +16003,10 @@ var TeamService = (() => {
   let _classSuper = TypertRemoteService;
   let _instanceExtraInitializers = [];
   let _remoteView_decorators;
-  let _remoteCreateTask_decorators;
-  let _remoteUpdateTask_decorators;
   return class TeamService extends _classSuper {
     static {
       const _metadata = typeof Symbol === "function" && Symbol.metadata ? Object.create(_classSuper[Symbol.metadata] ?? null) : void 0;
       _remoteView_decorators = [Remote("view")];
-      _remoteCreateTask_decorators = [Remote("createTask")];
-      _remoteUpdateTask_decorators = [Remote("updateTask")];
       __esDecorate(this, null, _remoteView_decorators, {
         kind: "method",
         name: "remoteView",
@@ -16021,28 +16018,6 @@ var TeamService = (() => {
         },
         metadata: _metadata
       }, null, _instanceExtraInitializers);
-      __esDecorate(this, null, _remoteCreateTask_decorators, {
-        kind: "method",
-        name: "remoteCreateTask",
-        static: false,
-        private: false,
-        access: {
-          has: (obj) => "remoteCreateTask" in obj,
-          get: (obj) => obj.remoteCreateTask
-        },
-        metadata: _metadata
-      }, null, _instanceExtraInitializers);
-      __esDecorate(this, null, _remoteUpdateTask_decorators, {
-        kind: "method",
-        name: "remoteUpdateTask",
-        static: false,
-        private: false,
-        access: {
-          has: (obj) => "remoteUpdateTask" in obj,
-          get: (obj) => obj.remoteUpdateTask
-        },
-        metadata: _metadata
-      }, null, _instanceExtraInitializers);
       if (_metadata) Object.defineProperty(this, Symbol.metadata, {
         enumerable: true,
         configurable: true,
@@ -16219,42 +16194,6 @@ var TeamService = (() => {
         tasks: this.listTasks(agent)
       };
     }
-    /**
-    * Create one shared task through the generated Remote API.
-    * @param agent - exact live Team member creating the task.
-    * @param request - task text, blockers, and advisory write scopes.
-    * @returns the revision-one task or a typed Team rejection.
-    */
-    remoteCreateTask(agent, request) {
-      return this.taskMutationResult(this.createTask(agent, request));
-    }
-    /**
-    * Apply one task mutation and preserve Team rejections as business results.
-    * @param agent - exact live Team member authorizing the mutation.
-    * @param request - task identity, expected revision, action, and action fields.
-    * @returns the committed task or a typed Team rejection.
-    */
-    remoteUpdateTask(agent, request) {
-      return this.taskMutationResult(this.updateTask(agent, request));
-    }
-    /** Preserve Team task rejections while allowing unexpected failures to reject the Remote call. */
-    async taskMutationResult(operation) {
-      try {
-        return {
-          ok: true,
-          value: await operation
-        };
-      } catch (error51) {
-        if (!(error51 instanceof TeamError)) throw error51;
-        return {
-          ok: false,
-          error: {
-            code: error51.code === "TEAM_TASK_STALE_REVISION" ? "team-task-conflict" : "team-rejected",
-            message: error51.message
-          }
-        };
-      }
-    }
     /** Queue one contained recovery pass after publication has unwound. */
     scheduleRecovery(agent) {
       queueMicrotask(() => {

package/dist/assets/presets/augmentcode/agent.cordis.yml

@@ -16,8 +16,9 @@
 - id: persona
   name: '@deepseek-ai/dsh-persona'
   config:
-    text: |
-      You are an Augment-style coding agent powered by the {{model}} model. Your working directory is {{cwd}}.
+    suffix: Your working directory is {{cwd}}.
+    prefix: |-
+      You are an Augment-style coding agent powered by the {{model}} model.
 
       # Output
 

package/dist/assets/presets/freecodego/agent.cordis.yml

@@ -15,8 +15,9 @@
 - id: persona
   name: '@deepseek-ai/dsh-persona'
   config:
-    text: |
-      You are a coding agent powered by the {{model}} model. Your working directory is {{cwd}}.
+    suffix: Your working directory is {{cwd}}.
+    prefix: |-
+      You are a coding agent powered by the {{model}} model.
 
       # Voice
 

package/dist/auto-review.js

@@ -1,5 +1,5 @@
 // packages/experimental/auto-review/lib/index.js
-import { BlockAssembler, createUserMessage } from "@deepseek-ai/dsh-llm";
+import { BlockAssembler } from "@deepseek-ai/dsh-llm";
 import { deepFreeze } from "@deepseek-ai/dsh-util-values";
 import { AUTO_PRESET } from "@deepseek-ai/dsh-permission-presets";
 import { RUN_CODE_NAME } from "@deepseek-ai/dsh-tools";
@@ -71,7 +71,7 @@ function isProjectInstruction(source) {
   return source.kind === "agent-instructions";
 }
 function isCheckpoint(source) {
-  return source.kind === "plugin" && source.plugin === "compact";
+  return source.kind === "compact-checkpoint";
 }
 function isDirectParentInstruction(source, parentSession) {
   return parentSession !== void 0 && source.kind === "agent-message" && source.senderSessionId === parentSession;
@@ -96,7 +96,7 @@ function textRole(source, seq, initialPromptSeq, parentSession) {
   return "fact";
 }
 function filteredUserEntries(seq, source, content, initialPromptSeq, parentSession) {
-  return content.filter((block) => block.type !== "tool-result").map((block) => ({
+  return content.map((block) => ({
     kind: "user-message",
     role: block.type === "text" ? textRole(source, seq, initialPromptSeq, parentSession) : "fact",
     source,
@@ -214,7 +214,7 @@ function snapshotAutoReview(agent, exec) {
     if (event.type === "user/message") {
       if (event.data.source.kind === "tool") continue;
       if (isProjectInstruction(event.data.source)) {
-        const content = event.data.content.filter((block) => block.type !== "tool-result");
+        const content = event.data.content;
         if (content.length > 0) projectInstructions.push({
           kind: "user-message",
           role: "constraint",
@@ -353,16 +353,13 @@ async function classifyRisk(ctx, agent, exec, signal) {
     provider: snapshot.provider,
     model: snapshot.model,
     system: REVIEW_POLICY,
-    messages: [createUserMessage({
+    messages: [{
+      role: "user",
       content: [{
         type: "text",
         text: reviewUserText(snapshot)
-      }],
-      source: {
-        kind: "plugin",
-        plugin: "dsh-experimental-auto-review"
-      }
-    })],
+      }]
+    }],
     temperature: 0,
     signal
   });