@hifullmoon/aicommit 2.0.1 → 2.2.0

package/docs/provider-compatibility.md

@@ -1,19 +1,18 @@
 # Provider compatibility / Provider 兼容表
 
-Provider presets choose setup defaults; adapters own request/response dialects; the core owns Git state, user interaction, HTTPS enforcement, retry, timeout, authorization, and machine output. Adding a compatible preset or an extension adapter does not modify the core Git/interaction flow.
+Built-in setup defaults choose an adapter and endpoint; adapters own request/response dialects; the core owns Git state, user interaction, HTTPS enforcement, retry, timeout, authorization, and machine output.
 
-Provider preset 只选择 setup 默认值；adapter 负责请求/响应方言；核心负责 Git 状态、用户交互、HTTPS、重试、超时、鉴权与机器输出。新增兼容 preset 或 extension adapter 无需修改核心 Git/交互流程。
+内置 setup 默认值负责选择 adapter 和端点；adapter 负责请求/响应方言；核心负责 Git 状态、用户交互、HTTPS、重试、超时、鉴权与机器输出。
 
-| Provider / adapter           | Endpoint and auth / 端点与鉴权                                                 | Streaming / 流式                                                                | Reasoning / 推理                                                                        | Token and usage mapping / token 与 usage                                           | Notes / 说明                                                                                                 |
-| ---------------------------- | ------------------------------------------------------------------------------ | ------------------------------------------------------------------------------- | --------------------------------------------------------------------------------------- | ---------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------------------------------ |
-| OpenAI / `openai`            | Official HTTPS Chat Completions; Bearer key / 官方 HTTPS；Bearer key           | SSE                                                                             | Native for recognized `o*`/`gpt-5*`; model-dependent otherwise / 已识别推理模型原生支持 | `max_completion_tokens` for reasoning models, otherwise `max_tokens`; OpenAI usage | Unsupported effort is rejected locally for known model generations / 已知模型不支持的 effort 会本地拒绝      |
-| OpenRouter / `openrouter`    | OpenRouter HTTPS; Bearer key; `X-Title: aicommit`                              | SSE                                                                             | `reasoning.effort`                                                                      | `max_tokens`; OpenAI-style usage                                                   | Model IDs commonly include vendor prefix / model 通常含厂商前缀                                              |
-| DeepSeek / `deepseek`        | DeepSeek HTTPS; Bearer key                                                     | SSE                                                                             | `thinking.type` plus normalized effort / thinking 与归一 effort                         | `max_tokens`; compatible usage                                                     | `medium`/`xhigh` map to supported high behavior where required / 必要时映射为 high                           |
-| MiniMax / `minimax`          | MiniMax HTTPS; Bearer key                                                      | SSE                                                                             | `reasoning_split` and thinking switch                                                   | `max_tokens`; compatible usage                                                     | Adapter removes conflicting switches before send / 发送前移除冲突开关                                        |
-| Kimi Code / `custom`         | Kimi Code OpenAI-compatible HTTPS; Bearer key / Kimi Code OpenAI 兼容 HTTPS    | SSE                                                                             | Server default; no vendor fields injected / 使用服务端默认值，不注入厂商字段            | `max_tokens`; OpenAI-style usage                                                   | Bundled preset uses `kimi-for-coding`; membership and Platform keys are distinct / 会员与开放平台 key 不通用 |
-| Ollama native / `ollama`     | Loopback `/api/chat` or `/api/generate`; normally keyless HTTP                 | Complete JSON in v1; native NDJSON streaming is not consumed / v1 使用完整 JSON | Native `think` boolean                                                                  | `options.num_predict`; `prompt_eval_count` + `eval_count`                          | OpenAI-compatible `/v1/chat/completions` uses the compatible shape instead / `/v1` 使用兼容方言              |
-| Custom compatible / `custom` | HTTPS remote or loopback HTTP; optional core Bearer key                        | SSE when endpoint supports Chat Completions events                              | No vendor fields by default; explicit `enabledBody`/`disabledBody` / 默认不注入厂商字段 | `max_tokens`; common OpenAI/Anthropic/Ollama usage fields normalized               | Validate the endpoint before trusting it with code or credentials / 发送代码前验证端点                       |
-| Extension / `extension:<id>` | Core HTTPS/loopback transport and optional Bearer only / 核心传输与可选 Bearer | Declared conservative in API v1 / v1 保守声明                                   | Adapter operation can transform non-secret reasoning config                             | Adapter maps body and normalized response                                          | No custom credential scheme or headers; credential-like body fields rejected / 不支持自定义鉴权或 header     |
+| Provider / adapter           | Endpoint and auth / 端点与鉴权                                              | Streaming / 流式                                                                | Reasoning / 推理                                                                        | Token and usage mapping / token 与 usage                                           | Notes / 说明                                                                                                 |
+| ---------------------------- | --------------------------------------------------------------------------- | ------------------------------------------------------------------------------- | --------------------------------------------------------------------------------------- | ---------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------------------------------ |
+| OpenAI / `openai`            | Official HTTPS Chat Completions; Bearer key / 官方 HTTPS；Bearer key        | SSE                                                                             | Native for recognized `o*`/`gpt-5*`; model-dependent otherwise / 已识别推理模型原生支持 | `max_completion_tokens` for reasoning models, otherwise `max_tokens`; OpenAI usage | Unsupported effort is rejected locally for known model generations / 已知模型不支持的 effort 会本地拒绝      |
+| OpenRouter / `openrouter`    | OpenRouter HTTPS; Bearer key; `X-Title: aicommit`                           | SSE                                                                             | `reasoning.effort`                                                                      | `max_tokens`; OpenAI-style usage                                                   | Model IDs commonly include vendor prefix / model 通常含厂商前缀                                              |
+| DeepSeek / `deepseek`        | DeepSeek HTTPS; Bearer key                                                  | SSE                                                                             | `thinking.type` plus normalized effort / thinking 与归一 effort                         | `max_tokens`; compatible usage                                                     | `medium`/`xhigh` map to supported high behavior where required / 必要时映射为 high                           |
+| MiniMax / `minimax`          | MiniMax HTTPS; Bearer key                                                   | SSE                                                                             | `reasoning_split` and thinking switch                                                   | `max_tokens`; compatible usage                                                     | Adapter removes conflicting switches before send / 发送前移除冲突开关                                        |
+| Kimi Code / `custom`         | Kimi Code OpenAI-compatible HTTPS; Bearer key / Kimi Code OpenAI 兼容 HTTPS | SSE                                                                             | Server default; no vendor fields injected / 使用服务端默认值，不注入厂商字段            | `max_tokens`; OpenAI-style usage                                                   | Bundled preset uses `kimi-for-coding`; membership and Platform keys are distinct / 会员与开放平台 key 不通用 |
+| Ollama native / `ollama`     | Loopback `/api/chat` or `/api/generate`; normally keyless HTTP              | Complete JSON in v1; native NDJSON streaming is not consumed / v1 使用完整 JSON | Native `think` boolean                                                                  | `options.num_predict`; `prompt_eval_count` + `eval_count`                          | OpenAI-compatible `/v1/chat/completions` uses the compatible shape instead / `/v1` 使用兼容方言              |
+| Custom compatible / `custom` | HTTPS remote or loopback HTTP; optional core Bearer key                     | SSE when endpoint supports Chat Completions events                              | No vendor fields by default; explicit `enabledBody`/`disabledBody` / 默认不注入厂商字段 | `max_tokens`; common OpenAI/Anthropic/Ollama usage fields normalized               | Validate the endpoint before trusting it with code or credentials / 发送代码前验证端点                       |
 
 ## Compatibility contract / 兼容契约
 
@@ -21,17 +20,6 @@ All built-in adapters return `content`, optional `reasoning`, normalized usage,
 
 所有内置 adapter 返回 `content`、可选 `reasoning`、标准 usage、finish reason、raw response、capability、attempts 与 latency。仅 429、部分 5xx、网络/响应中断会重试；鉴权、参数与安全错误不会重试。
 
-Preset compatibility is declared by core version range plus `adapterContract: 1`. Run:
+Use `aicommit doctor -p provider-name` to verify the selected adapter and live endpoint. Reuse a built-in adapter when only setup defaults change, and use `custom` for an OpenAI-compatible endpoint with optional body switches. A protocol requiring a different transport, streaming parser, or credential scheme needs core support.
 
-Preset 兼容性由核心版本范围与 `adapterContract: 1` 共同声明：
-
-```bash
-aicommit preset show
-aicommit preset validate --file=provider-presets.json
-aicommit preset install --file=provider-presets.json
-aicommit doctor -p provider-name
-```
-
-Use an existing built-in adapter when only setup defaults change. Use `custom` for an OpenAI-compatible endpoint with optional body switches. Use a reviewed `providerAdapter` extension only when the JSON request/response dialect differs and core Bearer authorization is sufficient. A protocol requiring different transport, streaming parser, or credential scheme is not compatible with extension API v1 and must not be disguised as a preset.
-
-若只改变 setup 默认值，应复用内置 adapter；OpenAI-compatible endpoint 及少量 body 开关使用 `custom`；仅当 JSON 请求/响应方言不同且核心 Bearer 鉴权足够时，才使用已审查的 `providerAdapter` 扩展。若协议需要不同传输、流解析或鉴权方案，则不兼容扩展 API v1，不能伪装成 preset。
+使用 `aicommit doctor -p provider-name` 校验所选 adapter 与在线端点。若只改变 setup 默认值，应复用内置 adapter；OpenAI-compatible endpoint 及少量 body 开关使用 `custom`。需要不同传输、流解析或鉴权方案的协议必须由核心直接支持。

package/docs/team-policy.md

@@ -46,7 +46,7 @@ Both paths call the same validator. `--output=json` returns the effective policy
 
 1. Move Conventional Commit types, scope rules, subject length, body rules, breaking-change handling, and language out of free-form `prompt` text and into `.aicommit.policy.json`.
 2. Declare every field instead of relying on personal defaults. Start with optional scopes and narrow the values only after current history has been sampled.
-3. Keep provider credentials in `~/.aicommit.config.json` or environment variables; never copy them into the repository policy.
+3. Keep provider credentials in `~/.aicommit/config.json` or environment variables; never copy them into the repository policy.
 4. If commitlint already defines `type-enum`, `scope-enum`, `subject-max-length`, or `header-max-length`, keep that file committed. AICommit reads recognized scalar values as data and never executes the config.
 5. Run the local hook and CI example on the same known-good and known-bad messages. Their `policyFingerprint` and issue codes must match before making the gate required.
 
@@ -54,7 +54,7 @@ Both paths call the same validator. `--output=json` returns the effective policy
 
 1. 将 Conventional Commit 类型、scope 规则、标题长度、正文规则、破坏性变更和语言要求从自由文本 `prompt` 迁移到 `.aicommit.policy.json`。
 2. 明确声明所有字段，不依赖个人默认值。可先保留可选 scope，再根据现有提交历史逐步收紧取值。
-3. provider 凭据继续放在 `~/.aicommit.config.json` 或环境变量中，绝不要复制到仓库策略。
+3. provider 凭据继续放在 `~/.aicommit/config.json` 或环境变量中，绝不要复制到仓库策略。
 4. 如果 commitlint 已定义 `type-enum`、`scope-enum`、`subject-max-length` 或 `header-max-length`，继续提交该配置。AICommit 只按数据读取识别出的标量规则，不执行配置文件。
 5. 用同一组已知正确/错误消息分别运行本地 hook 与 CI 示例；在强制启用门禁前，确认两者的 `policyFingerprint` 和问题代码一致。
 

package/docs/troubleshooting.md

@@ -15,22 +15,19 @@ JSON mode keeps one machine object on stdout and diagnostics on stderr. The `err
 
 JSON 模式保证 stdout 只有一个机器对象，诊断进入 stderr。`error.category` 与进程退出码可稳定用于自动化。
 
-| Symptom / 症状                                                                  | Category / exit                 | Likely cause / 常见原因                                                                                                                             | Check and recovery / 检查与恢复                                                                                                                         |
-| ------------------------------------------------------------------------------- | ------------------------------- | --------------------------------------------------------------------------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------- |
-| Config fails before Git/provider access / 在 Git/provider 前配置失败            | `config` / `2`                  | Malformed JSON, invalid URL, unsupported option, missing credential / JSON 损坏、URL 或选项无效、凭据缺失                                           | `aicommit config path`; `aicommit config validate`; fix only the user-owned file shown                                                                  |
-| Project settings are ignored / 项目设置被忽略                                   | warning                         | Repository config tried to set endpoint, credential, extension, reasoning, metrics, or raise a ceiling / 仓库配置尝试设置连接、凭据、扩展或提高预算 | Move trusted connection settings to `~/.aicommit.config.json`; keep team rules in `.aicommit.policy.json`                                               |
-| Not a Git repository, conflict, empty index / 非仓库、冲突或 index 为空         | `git_state` / `3`               | Wrong directory or unsafe Git state / 目录错误或 Git 状态不安全                                                                                     | `git status`; `aicommit /absolute/repo/path`; resolve conflicts before retrying                                                                         |
-| DNS, connection, or timeout / DNS、连接或超时                                   | `network` / `4`                 | Endpoint unavailable, proxy/TLS issue, budget too short / endpoint 不可达、代理/TLS、超时过短                                                       | `aicommit doctor -p NAME`; verify HTTPS URL; raise user-owned `timeoutMs` only if expected                                                              |
-| HTTP authentication/rate/parameter failure / 鉴权、限流或参数失败               | `provider` / `5`                | Wrong key/model/body; non-retryable 4xx / key、model、body 错误                                                                                     | Check `apiKeyEnv`, model and compatibility table; authentication is never retried automatically                                                         |
-| Empty or malformed model reply / 空或畸形回复                                   | `response_format` / `6`         | Unsupported response dialect, token limit, policy/extension validator failure / 响应方言、token 限制、policy/扩展校验失败                           | Raise `maxTokens`, choose the matching adapter, inspect validator issue code; one correction is already attempted                                       |
-| `split run --scope=all --yes` stops before API call / split 非交互在 API 前停止 | `sensitive_data` / `7`          | Complete untracked scan found sensitive-looking data / 完整未跟踪扫描发现疑似敏感数据                                                               | Review/stage intended files explicitly; do not bypass without checking the actual content                                                               |
-| Commit aborts after generation / 生成后提交中止                                 | `concurrent_modification` / `8` | Index/worktree changed during the protected window / 受保护窗口中 index/worktree 被修改                                                             | Review `git status`, restore the intended snapshot, and generate again                                                                                  |
-| Split stopped after one or more commits / split 部分提交后停止                  | reported Git failure            | Hook, crash, SIGINT, or concurrent pending edit / hook、崩溃、中断或待处理文件变化                                                                  | Run `aicommit split resume`; if another Git workflow already replaced the transaction, use `aicommit split abort`（只删除恢复元数据，不改提交或工作区） |
-| Preset install/rollback fails / preset 安装或回滚失败                           | `config` / `2`                  | Schema/core range/adapter contract mismatch or no valid backup / schema、核心范围、adapter contract 不匹配或无备份                                  | `aicommit preset validate --file=...`; `aicommit preset show`; follow the compatibility guide                                                           |
-| Extension requires Node 20+ / 扩展要求 Node 20+                                 | `config` / `2`                  | Node 18 intentionally has no executable-extension fallback / Node 18 有意不降级运行扩展                                                             | Upgrade Node for extensions or remove user manifest entries; core CLI still supports Node 18                                                            |
-| Extension timeout or permission error / 扩展超时或权限错误                      | warning or `response_format`    | Extension exceeded its limit or tried denied filesystem/process access / 扩展超时或访问被拒资源                                                     | Review the extension, keep it single-file, raise `extensions.timeoutMs` cautiously; validator failures are fail-closed                                  |
-| npm provenance is absent or invalid / npm provenance 缺失或失败                 | npm audit failure               | Old npm CLI, non-trusted release, or wrong version / npm 过旧、非可信发布或版本错误                                                                 | Upgrade npm; run `npm audit signatures`; install only a version linked to the official workflow                                                         |
-
-If a failure remains, capture `aicommit doctor --output=json`, Node/Git versions, the error category, and redacted config sources. Never attach a diff, commit message, config file, API key, reasoning trace, extension source containing secrets, or credential-helper output to a public issue.
-
-若问题仍未解决，请记录 `aicommit doctor --output=json`、Node/Git 版本、错误分类与脱敏后的配置来源。不要在公开 issue 中附加 diff、commit message、配置文件、API key、reasoning、含秘密的扩展源码或 credential-helper 输出。
+| Symptom / 症状                                                                  | Category / exit                 | Likely cause / 常见原因                                                                                                         | Check and recovery / 检查与恢复                                                                                                                         |
+| ------------------------------------------------------------------------------- | ------------------------------- | ------------------------------------------------------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------- |
+| Config fails before Git/provider access / 在 Git/provider 前配置失败            | `config` / `2`                  | Malformed JSON, invalid URL, unsupported option, missing credential / JSON 损坏、URL 或选项无效、凭据缺失                       | `aicommit config path`; `aicommit config validate`; fix only the user-owned file shown                                                                  |
+| Project settings are ignored / 项目设置被忽略                                   | warning                         | Repository config tried to set endpoint, credential, reasoning, or raise a ceiling / 仓库配置尝试设置连接、凭据、推理或提高预算 | Move trusted connection settings to `~/.aicommit/config.json`; keep team rules in `.aicommit.policy.json`                                               |
+| Not a Git repository, conflict, empty index / 非仓库、冲突或 index 为空         | `git_state` / `3`               | Wrong directory or unsafe Git state / 目录错误或 Git 状态不安全                                                                 | `git status`; `aicommit /absolute/repo/path`; resolve conflicts before retrying                                                                         |
+| DNS, connection, or timeout / DNS、连接或超时                                   | `network` / `4`                 | Endpoint unavailable, proxy/TLS issue, budget too short / endpoint 不可达、代理/TLS、超时过短                                   | `aicommit doctor -p NAME`; verify HTTPS URL; raise user-owned `timeoutMs` only if expected                                                              |
+| HTTP authentication/rate/parameter failure / 鉴权、限流或参数失败               | `provider` / `5`                | Wrong key/model/body; non-retryable 4xx / key、model、body 错误                                                                 | Check `apiKeyEnv`, model and compatibility table; authentication is never retried automatically                                                         |
+| Empty or malformed model reply / 空或畸形回复                                   | `response_format` / `6`         | Unsupported response dialect, token limit, or policy validation failure / 响应方言、token 限制或 policy 校验失败                | Raise `maxTokens`, choose the matching adapter, inspect validator issue code; one correction is already attempted                                       |
+| `split run --scope=all --yes` stops before API call / split 非交互在 API 前停止 | `sensitive_data` / `7`          | Complete untracked scan found sensitive-looking data / 完整未跟踪扫描发现疑似敏感数据                                           | Review/stage intended files explicitly; do not bypass without checking the actual content                                                               |
+| Commit aborts after generation / 生成后提交中止                                 | `concurrent_modification` / `8` | Index/worktree changed during the protected window / 受保护窗口中 index/worktree 被修改                                         | Review `git status`, restore the intended snapshot, and generate again                                                                                  |
+| Split stopped after one or more commits / split 部分提交后停止                  | reported Git failure            | Hook, crash, SIGINT, or concurrent pending edit / hook、崩溃、中断或待处理文件变化                                              | Run `aicommit split resume`; if another Git workflow already replaced the transaction, use `aicommit split abort`（只删除恢复元数据，不改提交或工作区） |
+| npm provenance is absent or invalid / npm provenance 缺失或失败                 | npm audit failure               | Old npm CLI, non-trusted release, or wrong version / npm 过旧、非可信发布或版本错误                                             | Upgrade npm; run `npm audit signatures`; install only a version linked to the official workflow                                                         |
+
+If a failure remains, capture `aicommit doctor --output=json`, Node/Git versions, the error category, and redacted config sources. Never attach a diff, commit message, config file, API key, reasoning trace, or credential-helper output to a public issue.
+
+若问题仍未解决，请记录 `aicommit doctor --output=json`、Node/Git 版本、错误分类与脱敏后的配置来源。不要在公开 issue 中附加 diff、commit message、配置文件、API key、reasoning 或 credential-helper 输出。

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@hifullmoon/aicommit",
-  "version": "2.0.1",
+  "version": "2.2.0",
   "description": "Safe, local-first AI commit message generator for Git workflows",
   "type": "module",
   "bin": {

package/presets/provider-presets.json

@@ -1,7 +1,7 @@
 {
   "kind": "aicommit-provider-presets",
   "schemaVersion": 2,
-  "version": "2.0.0",
+  "version": "2.1.0",
   "compatibility": {
     "coreMinimum": "1.5.1",
     "coreMaximumExclusive": "3.0.0",
@@ -18,6 +18,18 @@
         "gpt-4o": {
           "label": "GPT-4o",
           "modelId": "gpt-4o"
+        },
+        "gpt-5.6-sol": {
+          "label": "GPT-5.6 Sol",
+          "modelId": "gpt-5.6-sol"
+        },
+        "gpt-5.6-terra": {
+          "label": "GPT-5.6 Terra",
+          "modelId": "gpt-5.6-terra"
+        },
+        "gpt-5.6-luna": {
+          "label": "GPT-5.6 Luna",
+          "modelId": "gpt-5.6-luna"
         }
       }
     },
@@ -31,6 +43,10 @@
         "chat": {
           "label": "DeepSeek Chat",
           "modelId": "deepseek-v4-flash"
+        },
+        "pro": {
+          "label": "DeepSeek V4 Pro",
+          "modelId": "deepseek-v4-pro"
         }
       }
     },
@@ -43,6 +59,42 @@
       "models": {
         "default": {
           "modelId": "openai/gpt-4o-mini"
+        },
+        "auto": {
+          "label": "OpenRouter Auto",
+          "modelId": "openrouter/auto"
+        },
+        "gpt-5.6-terra": {
+          "label": "GPT-5.6 Terra",
+          "modelId": "openai/gpt-5.6-terra"
+        },
+        "claude-sonnet-5": {
+          "label": "Claude Sonnet 5",
+          "modelId": "anthropic/claude-sonnet-5"
+        },
+        "gemini-3.7-flash": {
+          "label": "Gemini 3.7 Flash",
+          "modelId": "google/gemini-3.7-flash"
+        },
+        "deepseek-v4-pro": {
+          "label": "DeepSeek V4 Pro",
+          "modelId": "deepseek/deepseek-v4-pro"
+        },
+        "qwen3.8-max": {
+          "label": "Qwen3.8 Max",
+          "modelId": "qwen/qwen3.8-max"
+        },
+        "glm-5.3": {
+          "label": "GLM-5.3",
+          "modelId": "z-ai/glm-5.3"
+        },
+        "kimi-k3": {
+          "label": "Kimi K3",
+          "modelId": "moonshotai/kimi-k3"
+        },
+        "grok-4.6": {
+          "label": "Grok 4.6",
+          "modelId": "x-ai/grok-4.6"
         }
       }
     },
@@ -86,6 +138,14 @@
         "qwen": {
           "label": "Qwen 3 8B",
           "modelId": "qwen3:8b"
+        },
+        "deepseek": {
+          "label": "DeepSeek R1 8B",
+          "modelId": "deepseek-r1:8b"
+        },
+        "gpt-oss": {
+          "label": "GPT-OSS 20B",
+          "modelId": "gpt-oss:20b"
         }
       }
     }

package/src/api.js

@@ -8,7 +8,6 @@ import {
   validateCommitCandidate,
 } from './policy.js';
 import { encodeUntrustedData } from './trust.js';
-import { extensionHostFor, resolveProviderAdapter } from './extensions.js';
 
 // Default per-request timeout; overridable via the "timeoutMs" config key.
 const DEFAULT_TIMEOUT_MS = 120_000;
@@ -167,7 +166,7 @@ async function fetchWithRetry(apiUrl, init, timeout, configuredPolicy, consume)
 export async function requestGeneration(config, request) {
   secureEndpoint(config.apiUrl);
   const timeout = config.timeoutMs || DEFAULT_TIMEOUT_MS;
-  const adapter = await resolveProviderAdapter(config, getProviderAdapter);
+  const adapter = getProviderAdapter(config);
   const payload = await adapter.buildRequest({
     messages: request.messages,
     temperature: request.temperature,
@@ -370,7 +369,7 @@ function hitTokenLimit(data) {
 // happened. Disable it only for providers where we know the switch is valid;
 // unknown compatible endpoints keep their configured behavior.
 async function reasoningForFollowUp(config) {
-  const adapter = await resolveProviderAdapter(config, getProviderAdapter);
+  const adapter = getProviderAdapter(config);
   return adapter.reasoningForFollowUp(config.reasoning);
 }
 
@@ -660,23 +659,7 @@ export async function generateCommitMessage(
   // Validate against the versioned policy and give the provider exactly one
   // cheap correction attempt. The diff is never re-sent: the prior reply plus
   // concrete violations are sufficient to repair formatting and constraints.
-  const validate = async (candidate) => {
-    const builtIn = validateCommitCandidate(candidate, { policy, diff });
-    const host = extensionHostFor(config);
-    if (!host) return builtIn;
-    const extensionIssues = await host.validateMessage(candidate, policy);
-    const issues = [...builtIn.issues, ...extensionIssues];
-    const errors = issues.filter((item) => item.severity === 'error');
-    const warnings = issues.filter((item) => item.severity === 'warning');
-    return {
-      ...builtIn,
-      valid: errors.length === 0,
-      needsCorrection: errors.length > 0,
-      issues,
-      errors,
-      warnings,
-    };
-  };
+  const validate = (candidate) => validateCommitCandidate(candidate, { policy, diff });
 
   let validation = await validate(message);
   if (message.trim() && validation.needsCorrection) {

package/src/cli.js

@@ -14,28 +14,24 @@ function showHelp() {
   ${chalk.bold('Usage:')}
     ${chalk.dim('$')} aicommit [path] [options]
     ${chalk.dim('$')} aicommit setup
-    ${chalk.dim('$')} aicommit config <show|validate|path> [options]
-    ${chalk.dim('$')} aicommit policy <template|check> [options]
-    ${chalk.dim('$')} aicommit preset <show|validate|path|install|rollback> [options]
-    ${chalk.dim('$')} aicommit completion <bash|zsh|fish>
     ${chalk.dim('$')} aicommit split [run|plan|apply|resume|abort] [options]
 
-  ${chalk.bold('Commands:')}
+  ${chalk.bold('Everyday commands:')}
     setup                 Interactive configuration wizard
     doctor                Diagnose runtime, config, credentials, and connectivity
+    split                 Plan and create file-level logical commits
+
+  ${chalk.bold('Advanced commands:')}
     config show           Print the effective configuration with secrets redacted
     config validate       Parse, merge, and validate configuration without reading credentials
     config path           Print user, project, and team-policy paths
     policy template       Print a safe repository team-policy template
     policy check          Validate a message file or Git range with the effective team policy
-    preset                Inspect, validate, install, or roll back provider preset manifests
     completion            Generate Bash, Zsh, or Fish completion on stdout
-    split, split run      Plan and create logical commits
     split plan            Generate and export a fingerprinted JSON plan
     split apply           Validate and apply an exported JSON plan
     split resume          Resume the repository's unfinished transaction
     split abort           Discard recovery metadata; keep commits and changes
-    stats [action]        Show quality/cost trends; show, clear, enable, disable
 
   ${chalk.bold('Arguments:')}
     path                  Target directory (default: current directory)
@@ -46,7 +42,6 @@ function showHelp() {
     -l, --lang=<zh|en>    Commit message language (default: zh)
     -p, --provider=<name> Use the named provider from config "providers"
     -m, --model=<name>    Use a named model from the selected provider
-    --split-hunks         Experimental same-file hunk planning (default: off)
     --scope=<scope>       Scope for "split|split plan": staged, all
     --file=<path>         Split-plan artifact or commit-message file
     --range=<revision>    Git revision/range for "policy check" (default: HEAD)
@@ -58,35 +53,11 @@ function showHelp() {
     --debug               Print debug info (parsed args, final config, etc.)
 
   ${chalk.bold('Examples:')}
-    aicommit setup        Run the interactive configuration wizard
-    aicommit doctor       Check runtime, config, credentials, and connectivity
-    aicommit config show  Show effective redacted configuration
-    aicommit config validate --output=json  Validate configuration for CI
-    aicommit config path  Show user and project config locations
-    aicommit policy template > .aicommit.policy.json  Create a committable policy
-    aicommit policy check --file=.git/COMMIT_EDITMSG  Validate a local commit message
-    aicommit policy check --range=origin/main..HEAD --output=json  Validate a CI range
-    aicommit preset show  Show the active versioned provider preset manifest
-    aicommit preset install --file=provider-presets.json  Install a compatible manifest
-    aicommit preset rollback  Restore the previously installed manifest
-    aicommit completion zsh > _aicommit  Generate Zsh completion
-    aicommit stats        Show local acceptance, quality, latency, and token trends
     aicommit              Commit changes in current directory (Chinese)
     aicommit --lang=en    Generate English commit message
-    aicommit -p deepseek -m chat  Select a configured provider and model
-    aicommit split        Choose staged/all scope, then plan logical commits
     aicommit split --scope=staged  Split only the reviewed index snapshot
-    aicommit split --scope=all --yes  Split the complete working tree non-interactively
-    aicommit split --scope=staged --split-hunks  Split eligible text hunks
-    aicommit split plan --scope=staged --file=.aicommit-plan.json --yes
-    aicommit split apply --file=.aicommit-plan.json --yes
-    aicommit split resume --yes  Resume pending groups from the checkpoint
-    aicommit split abort --yes  Discard a stale checkpoint without changing Git state
-    aicommit --reasoning=low  Stream reasoning; Ctrl+O expands/collapses it
     aicommit --dry-run    Review a generated message without committing
     aicommit --yes        Commit already staged changes without prompts
-    aicommit --yes --output=json  Emit one machine-readable result on stdout
-    aicommit /path/to    Commit changes in the specified directory
 `);
 }
 
@@ -115,7 +86,6 @@ function parsedDefaults(overrides = {}) {
     output: 'text',
     debug: false,
     split: null,
-    splitHunks: false,
     splitCommand: null,
     splitPlanFile: null,
     dryRun: false,
@@ -126,10 +96,7 @@ function parsedDefaults(overrides = {}) {
     policyAction: null,
     policyMessageFile: null,
     policyRange: null,
-    presetAction: null,
-    presetFile: null,
     completionShell: null,
-    statsAction: null,
     help: false,
     version: false,
     ...overrides,
@@ -137,6 +104,13 @@ function parsedDefaults(overrides = {}) {
 }
 
 export function parseArgs(args = process.argv.slice(2)) {
+  if (args[0] === 'stats' || args[0] === 'preset') {
+    throw fail(
+      ERROR_CATEGORIES.CONFIG,
+      `The "${args[0]}" command was removed to keep aicommit focused on commit generation.`,
+    );
+  }
+
   // "setup" is a standalone subcommand — it doesn't combine with the
   // commit-flow options, so short-circuit before parsing them.
   if (args[0] === 'setup') {
@@ -149,15 +123,6 @@ export function parseArgs(args = process.argv.slice(2)) {
     return parsedDefaults({ setup: true });
   }
 
-  if (args[0] === 'stats') {
-    const action = args[1] || 'show';
-    const allowed = ['show', 'clear', 'enable', 'disable'];
-    if (args.length > 2 || !allowed.includes(action)) {
-      throw fail(ERROR_CATEGORIES.CONFIG, `stats accepts one action: ${allowed.join(', ')}.`);
-    }
-    return parsedDefaults({ statsAction: action });
-  }
-
   if (args[0] === 'completion') {
     const shell = args[1];
     if (args.length !== 2 || !['bash', 'zsh', 'fish'].includes(shell)) {
@@ -182,18 +147,6 @@ export function parseArgs(args = process.argv.slice(2)) {
     args = args.slice(2);
   }
 
-  let presetAction = null;
-  if (args[0] === 'preset') {
-    presetAction = args[1];
-    if (!['show', 'validate', 'path', 'install', 'rollback'].includes(presetAction)) {
-      throw fail(
-        ERROR_CATEGORIES.CONFIG,
-        'preset requires one action: show, validate, path, install, or rollback.',
-      );
-    }
-    args = args.slice(2);
-  }
-
   let configAction = null;
   if (args[0] === 'config') {
     configAction = args[1];
@@ -235,11 +188,9 @@ export function parseArgs(args = process.argv.slice(2)) {
   let output = 'text';
   let debug = false;
   let split = null;
-  let splitHunks = false;
   let splitPlanFile = null;
   let policyMessageFile = null;
   let policyRange = null;
-  let presetFile = null;
   let splitScopeOption = false;
   let dryRun = false;
   let yes = false;
@@ -275,11 +226,6 @@ export function parseArgs(args = process.argv.slice(2)) {
       continue;
     }
 
-    if (arg === '--split-hunks') {
-      splitHunks = true;
-      continue;
-    }
-
     if (arg === '--scope') {
       splitScopeOption = true;
       split = takeValue(args, i, arg, 'staged|all');
@@ -302,7 +248,6 @@ export function parseArgs(args = process.argv.slice(2)) {
     if (arg === '--file') {
       const value = takeValue(args, i, arg, 'path');
       if (policyAction) policyMessageFile = value;
-      else if (presetAction) presetFile = value;
       else splitPlanFile = value;
       i++;
       continue;
@@ -312,7 +257,6 @@ export function parseArgs(args = process.argv.slice(2)) {
       const value = arg.slice('--file='.length);
       if (!value) throw fail(ERROR_CATEGORIES.CONFIG, 'Missing value for --file.');
       if (policyAction) policyMessageFile = value;
-      else if (presetAction) presetFile = value;
       else splitPlanFile = value;
       continue;
     }
@@ -430,12 +374,6 @@ export function parseArgs(args = process.argv.slice(2)) {
   if (!['text', 'json'].includes(output)) {
     throw fail(ERROR_CATEGORIES.CONFIG, `Invalid output mode: "${output}". Use text or json.`);
   }
-  if (splitHunks && !['run', 'plan'].includes(splitCommand)) {
-    throw fail(
-      ERROR_CATEGORIES.CONFIG,
-      '--split-hunks is only valid with "aicommit split run" or "aicommit split plan".',
-    );
-  }
   if (splitCommand) {
     if (['plan', 'apply'].includes(splitCommand) && !splitPlanFile) {
       throw fail(ERROR_CATEGORIES.CONFIG, `split ${splitCommand} requires --file=<path>.`);
@@ -453,10 +391,10 @@ export function parseArgs(args = process.argv.slice(2)) {
       }
       if (splitCommand === 'plan') dryRun = true;
     } else if (splitCommand === 'apply') {
-      if (split || splitHunks) {
+      if (split) {
         throw fail(
           ERROR_CATEGORIES.CONFIG,
-          'split apply reads its scope from the plan; do not pass --scope or --split-hunks.',
+          'split apply reads its scope from the plan; do not pass --scope.',
         );
       }
       if (cliProvider || cliModel || cliLang || cliReasoning || dryRun) {
@@ -466,16 +404,7 @@ export function parseArgs(args = process.argv.slice(2)) {
         );
       }
     } else {
-      if (
-        splitPlanFile ||
-        split ||
-        splitHunks ||
-        cliProvider ||
-        cliModel ||
-        cliLang ||
-        cliReasoning ||
-        dryRun
-      ) {
+      if (splitPlanFile || split || cliProvider || cliModel || cliLang || cliReasoning || dryRun) {
         throw fail(
           ERROR_CATEGORIES.CONFIG,
           `split ${splitCommand} accepts only an optional path, --yes, --output, and --debug.`,
@@ -494,14 +423,7 @@ export function parseArgs(args = process.argv.slice(2)) {
   }
   if (
     configAction &&
-    (cliLang ||
-      cliReasoning ||
-      split ||
-      splitHunks ||
-      splitCommand ||
-      splitPlanFile ||
-      dryRun ||
-      yes)
+    (cliLang || cliReasoning || split || splitCommand || splitPlanFile || dryRun || yes)
   ) {
     throw fail(
       ERROR_CATEGORIES.CONFIG,
@@ -513,15 +435,7 @@ export function parseArgs(args = process.argv.slice(2)) {
   }
   if (
     policyAction &&
-    (cliProvider ||
-      cliModel ||
-      cliLang ||
-      cliReasoning ||
-      split ||
-      splitHunks ||
-      splitCommand ||
-      dryRun ||
-      yes)
+    (cliProvider || cliModel || cliLang || cliReasoning || split || splitCommand || dryRun || yes)
   ) {
     throw fail(
       ERROR_CATEGORIES.CONFIG,
@@ -534,30 +448,6 @@ export function parseArgs(args = process.argv.slice(2)) {
   if (policyMessageFile && policyRange) {
     throw fail(ERROR_CATEGORIES.CONFIG, 'policy check accepts either --file or --range, not both.');
   }
-  if (
-    presetAction &&
-    (targetPath ||
-      cliProvider ||
-      cliModel ||
-      cliLang ||
-      cliReasoning ||
-      split ||
-      splitHunks ||
-      splitCommand ||
-      dryRun ||
-      yes)
-  ) {
-    throw fail(
-      ERROR_CATEGORIES.CONFIG,
-      'preset accepts only --file (validate/install), --output, and --debug.',
-    );
-  }
-  if (presetAction === 'install' && !presetFile) {
-    throw fail(ERROR_CATEGORIES.CONFIG, 'preset install requires --file=<manifest>.');
-  }
-  if (presetFile && !['validate', 'install'].includes(presetAction)) {
-    throw fail(ERROR_CATEGORIES.CONFIG, '--file is only valid with preset validate or install.');
-  }
   if (
     doctor &&
     (targetPath || cliLang || cliReasoning || split || splitPlanFile || dryRun || yes)
@@ -577,7 +467,6 @@ export function parseArgs(args = process.argv.slice(2)) {
     output,
     debug,
     split,
-    splitHunks,
     splitCommand,
     splitPlanFile,
     dryRun,
@@ -588,10 +477,7 @@ export function parseArgs(args = process.argv.slice(2)) {
     policyAction,
     policyMessageFile,
     policyRange,
-    presetAction,
-    presetFile,
     completionShell: null,
-    statsAction: null,
     help: false,
     version: false,
   };