golem15/msd-core
3
0
Fork 0
Code Issues Pull Requests Actions 1 Packages Projects Releases Wiki Activity

feat(#1014): add Claude Fable 5 model config (#1015)

Browse Source 
(cherry picked from commit 76f42ddb4b)
This commit is contained in:
Colin Johnson
2026-06-10 20:32:26 -04:00
committed by github-actions[bot]
parent 8c8c6e207a
commit 6d9d0e638a
9 changed files with 78 additions and 29 deletions
Download Patch File Download Diff File Expand all files Collapse all files

6
docs/CONFIGURATION.md
View File

@@ -144,7 +144,7 @@ GSD stores project settings in `.planning/config.json`. Created during `/gsd-new
| `model_profile` | enum | `quality`, `balanced`, `budget`, `adaptive`, `inherit` | `balanced` | Model tier for each agent (see [Model Profiles](#model-profiles)). `adaptive` was added per [#1713](https://github.com/open-gsd/gsd-core/issues/1713) / [#1806](https://github.com/open-gsd/gsd-core/issues/1806) and resolves the same way as the other tiers under runtime-aware profiles. |
| `runtime` | string | `claude`, `codex`, or any string | (none) | Active runtime for [runtime-aware profile resolution](#runtime-aware-profiles-2517). When set, profile tiers (opus/sonnet/haiku) resolve to runtime-native model IDs. Today only the Codex install path emits per-agent model IDs from this resolver; other runtimes (`opencode`, `gemini`, `qwen`, `copilot`, …) consume the resolver at spawn time and gain dedicated install-path support in [#2612](https://github.com/open-gsd/gsd-core/issues/2612). When unset (default), behavior is unchanged from prior versions. Added in v1.39 |
| `model_profile_overrides.<runtime>.<tier>` | string \| object | per-runtime tier override | (none) | Override the runtime-aware tier mapping for a specific `(runtime, tier)`. Tier is one of `opus`, `sonnet`, `haiku`. Value is either a model ID string (e.g. `"gpt-5-pro"`) or `{ model, reasoning_effort }`. See [Runtime-Aware Profiles](#runtime-aware-profiles-2517). Added in v1.39 |
| `model_policy.provider` | string | `openai`, `anthropic`, `google`, `qwen`, `generic` | (none) | Declares the model provider. Known providers (`openai`, `anthropic`, `google`, `qwen`) unlock catalog-backed presets. `generic` treats all model IDs as opaque strings — no prefix inference, no reasoning-effort defaults. `model_policy.runtime_tiers` resolves before legacy `model_profile_overrides`. See [Model Policy Presets](#model-policy-presets-model_policy--added-in-v142). Added in v1.42 ([#49](https://github.com/open-gsd/gsd-core/issues/49)) |
| `model_policy.provider` | string | `openai`, `anthropic`, `anthropic-fable`, `google`, `qwen`, `generic` | (none) | Declares the model provider. Known providers (`openai`, `anthropic`, `anthropic-fable`, `google`, `qwen`) unlock catalog-backed presets. `generic` treats all model IDs as opaque strings — no prefix inference, no reasoning-effort defaults. `model_policy.runtime_tiers` resolves before legacy `model_profile_overrides`. See [Model Policy Presets](#model-policy-presets-model_policy--added-in-v142). Added in v1.42 ([#49](https://github.com/open-gsd/gsd-core/issues/49)) |
| `model_policy.budget` | enum | `high`, `medium`, `low` | (none) | Selects a budget tier when using a known provider. GSD materializes the matching catalog preset into explicit tier mappings at resolve time. Ignored when `provider` is `generic` or `custom`. Added in v1.42 ([#49](https://github.com/open-gsd/gsd-core/issues/49)) |
| `model_policy.high` | string | model ID | (none) | High-cost tier model ID for `generic`/`custom` provider. Used when `provider: "generic"` or `"custom"`. Added in v1.42 ([#49](https://github.com/open-gsd/gsd-core/issues/49)) |
| `model_policy.medium` | string | model ID | (none) | Medium-cost tier model ID for `generic`/`custom` provider. Added in v1.42 ([#49](https://github.com/open-gsd/gsd-core/issues/49)) |
@@ -159,7 +159,7 @@ GSD stores project settings in `.planning/config.json`. Created during `/gsd-new
| `project_code` | string | any short string | (none) | Prefix for phase directory names (e.g., `"ABC"` produces `ABC-01-setup/`). Added in v1.31 |
| `phase_id_convention` | enum | `"milestone-prefixed"`, `null` | `null` | Phase ID naming convention. `null` = legacy numeric IDs (`Phase 1`, `Phase 2`). `"milestone-prefixed"` = globally unique IDs that encode the enclosing milestone (`Phase 1-01`, `Phase 1-02`). Run `gsd-tools roadmap upgrade --convention milestone-prefixed` to migrate an existing ROADMAP.md. |
| `response_language` | string | language code | (none) | Language for agent responses (e.g., `"pt"`, `"ko"`, `"ja"`). Propagates to all spawned agents for cross-phase language consistency. Added in v1.32 |
| `context_window` | number | any integer | `200000` | Context window size in tokens. Set `1000000` for 1M-context models (e.g., `claude-opus-4-7[1m]`). Values `>= 500000` enable adaptive context enrichment (full-body reads of prior SUMMARY.md, deeper anti-pattern reads). Configured via `/gsd-config --advanced`. |
| `context_window` | number | any integer | `200000` | Context window size in tokens. Set `1000000` for 1M-context models (e.g., `claude-fable-5`). Values `>= 500000` enable adaptive context enrichment (full-body reads of prior SUMMARY.md, deeper anti-pattern reads). Configured via `/gsd-config --advanced`. |
| `context_profile` | string | `dev`, `research`, `review` | (none) | Execution context preset that applies a pre-configured bundle of mode, model, and workflow settings for the current type of work. Added in v1.34 |
| `claude_md_path` | string | any file path | `./CLAUDE.md` | Custom output path for the generated CLAUDE.md file. Useful for monorepos or projects that need CLAUDE.md in a non-root location. Defaults to `./CLAUDE.md` at the project root. Added in v1.36 |
| `claude_md_assembly.mode` | enum | `embed`, `link` | `embed` | Controls how managed sections are written into CLAUDE.md. `embed` (default) inlines content between GSD markers. `link` writes `@.planning/<source-path>` instead — Claude Code expands the reference at runtime, reducing CLAUDE.md size by ~65% on typical projects. `link` only applies to sections that have a real source file; `workflow` and fallback sections always embed. Per-block overrides: `claude_md_assembly.blocks.<section>` (e.g. `claude_md_assembly.blocks.architecture: link`). Added in v1.38 |
@@ -1284,7 +1284,7 @@ Choose a provider and budget level via the settings workflow; GSD writes the can
}
```
Known providers: `openai`, `anthropic`, `google`, `qwen`. Budget levels: `high`, `medium`, `low`.
Known providers: `openai`, `anthropic`, `anthropic-fable`, `google`, `qwen`. Budget levels: `high`, `medium`, `low`. Use `anthropic` to keep the Opus 4.8-backed Claude preset, or `anthropic-fable` to opt into Claude Fable 5 for high-budget top-tier routing.
For advanced per-runtime control, `runtime_tiers` accepts explicit entries using the internal profile tier names (`opus`, `sonnet`, `haiku`):

2
docs/ja-JP/CONFIGURATION.md
View File

@@ -331,7 +331,7 @@ GSD が非 Claude ランタイム向けにインストールされると、イ
| 値 | 動作 | 使用場面 |
|----|------|---------|
| `false`（デフォルト） | Claude エイリアス（`opus`、`sonnet`、`haiku`）を返す | Claude Code + ネイティブ Anthropic API |
| `true` | エイリアスを完全な Claude モデル ID（`claude-opus-4-6`）にマッピング | 完全な ID が必要な API を使用する Claude Code |
| `true` | エイリアスを完全な Claude モデル ID（`claude-opus-4-8`）にマッピング | 完全な ID が必要な API を使用する Claude Code |
| `"omit"` | 空文字列を返す（ランタイムがデフォルトを選択） | 非 Claude ランタイム（Codex、OpenCode、Gemini CLI、Kilo） |
### プロファイルの設計思想

2
docs/ko-KR/CONFIGURATION.md
View File

@@ -331,7 +331,7 @@ quick 태스크 브랜칭 예시:
| 값 | 동작 | 사용 시점 |
|----|------|----------|
| `false` (기본값) | Claude 별칭 반환 (`opus`, `sonnet`, `haiku`) | 네이티브 Anthropic API를 사용하는 Claude Code |
| `true` | 별칭을 전체 Claude 모델 ID로 매핑 (`claude-opus-4-6`) | 전체 ID가 필요한 API를 사용하는 Claude Code |
| `true` | 별칭을 전체 Claude 모델 ID로 매핑 (`claude-opus-4-8`) | 전체 ID가 필요한 API를 사용하는 Claude Code |
| `"omit"` | 빈 문자열 반환 (런타임이 기본값 선택) | 비 Claude 런타임 (Codex, OpenCode, Gemini CLI, Kilo) |
### 프로필 철학

8
docs/pt-BR/CONFIGURATION.md
View File

@@ -141,7 +141,7 @@ O GSD armazena as configurações do projeto em `.planning/config.json`. Criado
| `model_profile` | enum | `quality`, `balanced`, `budget`, `adaptive`, `inherit` | `balanced` | Nível de modelo para cada agente (consulte [Perfis de Modelo](#model-profiles)). `adaptive` foi adicionado conforme [#1713](https://github.com/open-gsd/gsd-core/issues/1713) / [#1806](https://github.com/open-gsd/gsd-core/issues/1806) e resolve da mesma forma que os outros níveis em perfis com reconhecimento de runtime. |
| `runtime` | string | `claude`, `codex`, ou qualquer string | (nenhum) | Runtime ativo para [resolução de perfil com reconhecimento de runtime](#runtime-aware-profiles-2517). Quando definido, os níveis de perfil (opus/sonnet/haiku) resolvem para IDs de modelo nativos do runtime. Atualmente, apenas o caminho de instalação do Codex emite IDs de modelo por agente a partir deste resolvedor; outros runtimes (`opencode`, `gemini`, `qwen`, `copilot`, …) consomem o resolvedor no momento do spawn e ganham suporte a caminho de instalação dedicado em [#2612](https://github.com/open-gsd/gsd-core/issues/2612). Quando não definido (padrão), o comportamento não se altera em relação às versões anteriores. Adicionado na v1.39 |
| `model_profile_overrides.<runtime>.<tier>` | string \| object | substituição de nível por runtime | (nenhum) | Substitui o mapeamento de nível com reconhecimento de runtime para um `(runtime, tier)` específico. O nível é um de `opus`, `sonnet`, `haiku`. O valor é uma string de ID de modelo (por exemplo, `"gpt-5-pro"`) ou `{ model, reasoning_effort }`. Consulte [Perfis com Reconhecimento de Runtime](#runtime-aware-profiles-2517). Adicionado na v1.39 |
| `model_policy.provider` | string | `openai`, `anthropic`, `google`, `qwen`, `generic` | (nenhum) | Declara o provedor de modelo. Provedores conhecidos (`openai`, `anthropic`, `google`, `qwen`) desbloqueiam predefinições baseadas em catálogo. `generic` trata todos os IDs de modelo como strings opacas — sem inferência de prefixo, sem padrões de esforço de raciocínio. `model_policy.runtime_tiers` resolve antes do legado `model_profile_overrides`. Consulte [Predefinições de Política de Modelo](#model-policy-presets-model_policy--added-in-v142). Adicionado na v1.42 ([#49](https://github.com/open-gsd/gsd-core/issues/49)) |
| `model_policy.provider` | string | `openai`, `anthropic`, `anthropic-fable`, `google`, `qwen`, `generic` | (nenhum) | Declara o provedor de modelo. Provedores conhecidos (`openai`, `anthropic`, `anthropic-fable`, `google`, `qwen`) desbloqueiam predefinições baseadas em catálogo. `generic` trata todos os IDs de modelo como strings opacas — sem inferência de prefixo, sem padrões de esforço de raciocínio. `model_policy.runtime_tiers` resolve antes do legado `model_profile_overrides`. Consulte [Predefinições de Política de Modelo](#model-policy-presets-model_policy--added-in-v142). Adicionado na v1.42 ([#49](https://github.com/open-gsd/gsd-core/issues/49)) |
| `model_policy.budget` | enum | `high`, `medium`, `low` | (nenhum) | Seleciona um nível de orçamento ao usar um provedor conhecido. O GSD materializa a predefinição de catálogo correspondente em mapeamentos de nível explícitos no momento da resolução. Ignorado quando `provider` é `generic` ou `custom`. Adicionado na v1.42 ([#49](https://github.com/open-gsd/gsd-core/issues/49)) |
| `model_policy.high` | string | ID do modelo | (nenhum) | ID do modelo de nível de custo alto para provedor `generic`/`custom`. Usado quando `provider: "generic"` ou `"custom"`. Adicionado na v1.42 ([#49](https://github.com/open-gsd/gsd-core/issues/49)) |
| `model_policy.medium` | string | ID do modelo | (nenhum) | ID do modelo de nível de custo médio para provedor `generic`/`custom`. Adicionado na v1.42 ([#49](https://github.com/open-gsd/gsd-core/issues/49)) |
@@ -155,7 +155,7 @@ O GSD armazena as configurações do projeto em `.planning/config.json`. Criado
| `project_code` | string | qualquer string curta | (nenhum) | Prefixo para nomes de diretórios de fase (por exemplo, `"ABC"` produz `ABC-01-setup/`). Adicionado na v1.31 |
| `phase_id_convention` | enum | `"milestone-prefixed"`, `null` | `null` | Convenção de nomenclatura para IDs de fase. `null` = IDs numéricos legados (`Phase 1`, `Phase 2`). `"milestone-prefixed"` = IDs globalmente únicos que codificam o marco envolvente (`Phase 1-01`, `Phase 1-02`). Execute `gsd-tools roadmap upgrade --convention milestone-prefixed` para migrar um ROADMAP.md existente. |
| `response_language` | string | código de idioma | (nenhum) | Idioma para respostas dos agentes (por exemplo, `"pt"`, `"ko"`, `"ja"`). Propagado para todos os agentes gerados para consistência de idioma entre fases. Adicionado na v1.32 |
| `context_window` | number | qualquer inteiro | `200000` | Tamanho da janela de contexto em tokens. Defina `1000000` para modelos com contexto de 1M (por exemplo, `claude-opus-4-7[1m]`). Valores `>= 500000` habilitam enriquecimento adaptativo de contexto (leituras completas de SUMMARY.md anteriores, leituras mais profundas de antipadrões). Configurado via `/gsd-config --advanced`. |
| `context_window` | number | qualquer inteiro | `200000` | Tamanho da janela de contexto em tokens. Defina `1000000` para modelos com contexto de 1M (por exemplo, `claude-fable-5`). Valores `>= 500000` habilitam enriquecimento adaptativo de contexto (leituras completas de SUMMARY.md anteriores, leituras mais profundas de antipadrões). Configurado via `/gsd-config --advanced`. |
| `context_profile` | string | `dev`, `research`, `review` | (nenhum) | Predefinição de contexto de execução que aplica um conjunto pré-configurado de configurações de modo, modelo e fluxo de trabalho para o tipo atual de trabalho. Adicionado na v1.34 |
| `claude_md_path` | string | qualquer caminho de arquivo | `./CLAUDE.md` | Caminho de saída personalizado para o arquivo CLAUDE.md gerado. Útil para monorepos ou projetos que precisam do CLAUDE.md em um local fora da raiz. Padrão é `./CLAUDE.md` na raiz do projeto. Adicionado na v1.36 |
| `claude_md_assembly.mode` | enum | `embed`, `link` | `embed` | Controla como as seções gerenciadas são escritas no CLAUDE.md. `embed` (padrão) incorpora conteúdo entre marcadores GSD. `link` escreve `@.planning/<source-path>` — o Claude Code expande a referência em tempo de execução, reduzindo o tamanho do CLAUDE.md em ~65% em projetos típicos. `link` aplica-se apenas a seções que possuem um arquivo-fonte real; as seções `workflow` e fallback sempre são incorporadas. Substituições por bloco: `claude_md_assembly.blocks.<section>` (por exemplo `claude_md_assembly.blocks.architecture: link`). Adicionado na v1.38 |
@@ -994,7 +994,7 @@ O hint `reasoning_effort` por nível do catálogo de modelos é um campo legado
2. `effort.agent_overrides[<agent-id>]`
3. `effort.routing_tier_defaults[<light|standard|heavy>]`
4. `effort.default`
5. `"high"` (padrão universal do Anthropic Opus 4.8)
5. `"high"` (padrão universal do Claude)
```json
{
@@ -1229,7 +1229,7 @@ Escolha um provedor e nível de orçamento via o fluxo de configurações; o GSD
}
```
Provedores conhecidos: `openai`, `anthropic`, `google`, `qwen`. Níveis de orçamento: `high`, `medium`, `low`.
Provedores conhecidos: `openai`, `anthropic`, `anthropic-fable`, `google`, `qwen`. Níveis de orçamento: `high`, `medium`, `low`. Use `anthropic` para manter a predefinição Claude baseada em Opus 4.8, ou `anthropic-fable` para optar pelo Claude Fable 5 no roteamento de alto orçamento.
Para controle avançado por runtime, `runtime_tiers` aceita entradas explícitas usando os nomes internos de nível de perfil (`opus`, `sonnet`, `haiku`):

8
docs/zh-CN/CONFIGURATION.md
View File

@@ -141,7 +141,7 @@ GSD 将项目设置存储在 `.planning/config.json` 中。该文件在 `/gsd-ne
| `model_profile` | enum | `quality`, `balanced`, `budget`, `adaptive`, `inherit` | `balanced` | 每个 agent 的模型层级（参见[模型配置文件](#模型配置文件)）。`adaptive` 根据 [#1713](https://github.com/open-gsd/gsd-core/issues/1713) / [#1806](https://github.com/open-gsd/gsd-core/issues/1806) 添加，在运行时感知的配置文件下与其他层级以相同方式解析。 |
| `runtime` | string | `claude`, `codex` 或任意字符串 | （无） | [运行时感知配置文件解析](#运行时感知配置文件-2517)的活跃运行时。设置后，配置文件层级（opus/sonnet/haiku）解析为运行时原生模型 ID。目前仅 Codex 安装路径通过此解析器为每个 agent 生成模型 ID；其他运行时（`opencode`、`gemini`、`qwen`、`copilot` 等）在 spawn 时消费该解析器，并在 [#2612](https://github.com/open-gsd/gsd-core/issues/2612) 中获得专用安装路径支持。未设置时（默认），行为与之前版本相同。v1.39 新增 |
| `model_profile_overrides.<runtime>.<tier>` | string \| object | 按运行时的层级覆盖 | （无） | 覆盖特定 `(runtime, tier)` 的运行时感知层级映射。层级为 `opus`、`sonnet`、`haiku` 之一。值为模型 ID 字符串（如 `"gpt-5-pro"`）或 `{ model, reasoning_effort }`。参见[运行时感知配置文件](#运行时感知配置文件-2517)。v1.39 新增 |
| `model_policy.provider` | string | `openai`, `anthropic`, `google`, `qwen`, `generic` | （无） | 声明模型提供商。已知提供商（`openai`、`anthropic`、`google`、`qwen`）启用基于目录的预设。`generic` 将所有模型 ID 视为不透明字符串——无前缀推断，无推理努力默认值。`model_policy.runtime_tiers` 在旧版 `model_profile_overrides` 之前解析。参见[模型策略预设](#模型策略预设-model_policy--v142-新增)。v1.42 新增（[#49](https://github.com/open-gsd/gsd-core/issues/49)） |
| `model_policy.provider` | string | `openai`, `anthropic`, `anthropic-fable`, `google`, `qwen`, `generic` | （无） | 声明模型提供商。已知提供商（`openai`、`anthropic`、`anthropic-fable`、`google`、`qwen`）启用基于目录的预设。`generic` 将所有模型 ID 视为不透明字符串——无前缀推断，无推理努力默认值。`model_policy.runtime_tiers` 在旧版 `model_profile_overrides` 之前解析。参见[模型策略预设](#模型策略预设-model_policy--v142-新增)。v1.42 新增（[#49](https://github.com/open-gsd/gsd-core/issues/49)） |
| `model_policy.budget` | enum | `high`, `medium`, `low` | （无） | 使用已知提供商时选择预算层级。GSD 在解析时将匹配的目录预设具体化为显式层级映射。当 `provider` 为 `generic` 或 `custom` 时忽略。v1.42 新增（[#49](https://github.com/open-gsd/gsd-core/issues/49)） |
| `model_policy.high` | string | 模型 ID | （无） | `generic`/`custom` 提供商的高成本层级模型 ID。当 `provider: "generic"` 或 `"custom"` 时使用。v1.42 新增（[#49](https://github.com/open-gsd/gsd-core/issues/49)） |
| `model_policy.medium` | string | 模型 ID | （无） | `generic`/`custom` 提供商的中等成本层级模型 ID。v1.42 新增（[#49](https://github.com/open-gsd/gsd-core/issues/49)） |
@@ -155,7 +155,7 @@ GSD 将项目设置存储在 `.planning/config.json` 中。该文件在 `/gsd-ne
| `project_code` | string | 任意短字符串 | （无） | 阶段目录名称的前缀（如 `"ABC"` 生成 `ABC-01-setup/`）。v1.31 新增 |
| `phase_id_convention` | enum | `"milestone-prefixed"`, `null` | `null` | 阶段 ID 命名规范。`null` = 旧版数字 ID（`Phase 1`、`Phase 2`）。`"milestone-prefixed"` = 编码所属里程碑的全局唯一 ID（`Phase 1-01`、`Phase 1-02`）。运行 `gsd-tools roadmap upgrade --convention milestone-prefixed` 迁移现有 ROADMAP.md。 |
| `response_language` | string | 语言代码 | （无） | agent 响应语言（如 `"pt"`、`"ko"`、`"ja"`）。传播至所有派生 agent，实现跨阶段语言一致性。v1.32 新增 |
| `context_window` | number | 任意整数 | `200000` | 上下文窗口大小（token 数）。对于 1M 上下文模型（如 `claude-opus-4-7[1m]`），设置为 `1000000`。`>= 500000` 的值启用自适应上下文增强（完整读取之前的 SUMMARY.md，更深入的反模式读取）。通过 `/gsd-config --advanced` 配置。 |
| `context_window` | number | 任意整数 | `200000` | 上下文窗口大小（token 数）。对于 1M 上下文模型（如 `claude-fable-5`），设置为 `1000000`。`>= 500000` 的值启用自适应上下文增强（完整读取之前的 SUMMARY.md，更深入的反模式读取）。通过 `/gsd-config --advanced` 配置。 |
| `context_profile` | string | `dev`, `research`, `review` | （无） | 执行上下文预设，为当前工作类型应用预配置的模式、模型和工作流设置包。v1.34 新增 |
| `claude_md_path` | string | 任意文件路径 | `./CLAUDE.md` | 生成的 CLAUDE.md 文件的自定义输出路径。适用于需要将 CLAUDE.md 放在非根目录位置的 monorepo 或项目。默认为项目根目录下的 `./CLAUDE.md`。v1.36 新增 |
| `claude_md_assembly.mode` | enum | `embed`, `link` | `embed` | 控制如何将受管理的节写入 CLAUDE.md。`embed`（默认）在 GSD 标记之间内联内容。`link` 改为写入 `@.planning/<source-path>`——Claude Code 在运行时展开引用，在典型项目中将 CLAUDE.md 大小减少约 65%。`link` 仅适用于有真实源文件的节；`workflow` 和回退节始终嵌入。按块覆盖：`claude_md_assembly.blocks.<section>`（如 `claude_md_assembly.blocks.architecture: link`）。v1.38 新增 |
@@ -963,7 +963,7 @@ minimal < low < medium < high < xhigh < max
2. `effort.agent_overrides[<agent-id>]`
3. `effort.routing_tier_defaults[<light|standard|heavy>]`
4. `effort.default`
5. `"high"`（Anthropic Opus 4.8 通用默认值）
5. `"high"`（Claude 通用默认值）
```json
{
@@ -1198,7 +1198,7 @@ minimal < low < medium < high < xhigh < max
}
```
已知提供商：`openai`、`anthropic`、`google`、`qwen`。预算级别：`high`、`medium`、`low`。
已知提供商：`openai`、`anthropic`、`anthropic-fable`、`google`、`qwen`。预算级别：`high`、`medium`、`low`。使用 `anthropic` 保留基于 Opus 4.8 的 Claude 预设，或使用 `anthropic-fable` 在高预算路由中选择 Claude Fable 5。
对于高级的按运行时控制，`runtime_tiers` 接受使用内部配置文件层级名称（`opus`、`sonnet`、`haiku`）的显式条目：