chore(repo): reorganize tools/ directory

.github/workflows/ci.yml

@@ -114,4 +114,4 @@ jobs:
         shell: pwsh
         run: |
           git fetch origin main
-          tools/invariant-checks/run-checks.ps1
+          scripts/invariant-checks/run-checks.ps1

.gitignore

@@ -28,14 +28,9 @@ backup-*.db
 # OS metadata
 .DS_Store
 Thumbs.db
-tools/embedding-provider/__pycache__/
-
-tools/embedding-provider/__pycache__/
-
+examples/embedding-provider/__pycache__/
 
 .agents/
 .venv/
-spike/**/target/
 coverage_report.txt
 /releases/
-.kimi/

CHANGELOG.md

@@ -15,7 +15,7 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
   - Tantivy 一致性修复 — `repair_tantivy_consistency_at()` 启动时自动检测 orphan/missing 文档
   - 性能回归基线 — `test_keyword_search_latency_regression_1k` / `_10k`（profile-aware 阈值）
   - `TempStorageBackend` — 测试隔离后端，消除 `DEVBASE_DATA_DIR` 竞态
-- **Architecture Invariants CI 自动化** — `tools/invariant-checks/run-checks.ps1`
+- **Architecture Invariants CI 自动化** — `scripts/invariant-checks/run-checks.ps1`
   - G5 (RF-6)：diff-only 检测新增生产代码 `unwrap`/`expect`/`panic`（排除 `#[cfg(test)]`）
   - T11：`mcp/tools` 禁止直接调用 `rusqlite::Connection`
   - T12：`tui/render` 纯消费检查（禁止写入操作）
@@ -172,7 +172,7 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
   - 新增: `python`, `bun`, `zig`, `java`
   - `get_tool_version` 支持 stderr fallback (Java 输出到 stderr)
   - `fmt_version` 改进: Java 引号提取、Docker/Python 格式处理
-- **P5: 架构不变量自动化 CI** — `tools/invariant-checks/run-checks.ps1`
+- **P5: 架构不变量自动化 CI** — `scripts/invariant-checks/run-checks.ps1`
   - G5: diff-only 检测新增生产代码 unwrap/expect/panic（排除 `#[cfg(test)]`）
   - T11: 检测 `mcp/tools/*` 直接调用 `rusqlite::Connection`
   - T12: 检测 `tui/render/*` 写入操作
@@ -485,7 +485,7 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
   - Schema v13: `code_symbol_links` table (source_repo, source_symbol, target_repo, target_symbol, link_type, strength)
   - `src/symbol_links.rs`: `compute_similar_signature_links()` (Jaccard token overlap), `compute_co_located_links()` (same-file clustering)
   - `generate_and_save_links()`: persists links with ON CONFLICT IGNORE upsert
-- **External Embedding Provider** — Reference Python implementation in `tools/embedding-provider/`
+- **External Embedding Provider** — Reference Python implementation in `examples/embedding-provider/`
   - `index.py`: Ollama `/api/embeddings` client, batch generation, cross-platform registry DB path
   - Byte-compatible f32 little-endian serialization via `struct.pack`
   - CLI: `--repo-id`, `--model`, `--ollama-url`, `--batch-size`, `--force`

docs/AGENTS-full.md

@@ -292,7 +292,7 @@ done
 | 10 | OpLog 结构化 | Schema v12, OplogEventType 枚举, JSON metadata, duration_ms | `7aa2a65` |
 | 11 | 性能基准 | criterion benches: index_repo_full, cosine_similarity, extract_symbols, CMake | `8e0f236` |
 | 12 | 混合检索核心 | `search::hybrid.rs`: RRF 归并, keyword_search, hybrid_search_symbols | `7fca714` |
-| 13 | 外部 Embedding Provider | Python CLI `tools/embedding-provider/`, Ollama 批量生成, 字节兼容序列化 | `574fb96` |
+| 13 | 外部 Embedding Provider | Python CLI `examples/embedding-provider/`, Ollama 批量生成, 字节兼容序列化 | `574fb96` |
 | 14a | 跨 repo 语义聚合 | `cross_repo_search_symbols()` INTERSECT tag 过滤, `devkit_cross_repo_search` | `8e762c7` |
 | 14b | 知识覆盖报告 | `oplog_analytics.rs`: 表存在性容错, 覆盖度/健康度/活动流, `devkit_knowledge_report` | `869bcbf` |
 | 15a | 显式知识链接 | Schema v13 `code_symbol_links`, Jaccard 签名相似度, 同文件聚类, `devkit_related_symbols` | `d462209` |
@@ -453,7 +453,7 @@ v0.11.3 已交付（tagged）。v0.12.0-alpha 全部功能已完成，进入发
 **Embedding 状态**：
 - `code_embeddings`: **56,722** 行（37.0% 覆盖率），覆盖 10 个仓库
 - `skills.embedding`: 3 个 builtin skill 已有 384-dim 向量
-- 生成工具：`tools/embedding-provider/skills.py`（sentence-transformers `all-MiniLM-L6-v2`）
+- 生成工具：`examples/embedding-provider/skills.py`（sentence-transformers `all-MiniLM-L6-v2`）
 - 激活路径：启动 Ollama + `devbase index <repo>` 生成 embedding，或配置远程 provider 于 `config.toml [embedding]` 段
 
 ### 2026-05-04 索引性能实验记录

docs/theory/three-repositories.md

@@ -140,7 +140,7 @@ devbase 的当前架构已经**天然横跨三层**，只是没有明确形式
 | `search/hybrid` + `semantic_index` | Layer 1→2 桥梁 | ✅ 成熟：keyword + vector + RRF |
 | `mcp/` (31 tools) | Layer 3 (MCP 仓库) | ✅ 成熟：Stable/Beta/Experimental 三级 |
 | `vault/` + `skill-sync-prototype` | Layer 2 (Skill 仓库) | 🔄 原型：Vault→Skill 同步设计完成，待实现 |
-| `tools/embedding-provider/` | Layer 2→3 桥梁 | ✅ 可用：local.py 引擎无关 |
+| `examples/embedding-provider/` | Layer 2→3 桥梁 | ✅ 可用：local.py 引擎无关 |
 | **Skill Runtime** (`devbase skill` CLI) | Layer 2 (Skill 仓库) | ❌ 未开始：registry + executor + sandbox |
 | **Protocol Adapter** (MCP + OpenAPI 统一) | Layer 3 (MCP 仓库) | ❌ 未开始：ToolRegistry 模式移植 |
 

tools/embedding-provider-rs/README.md → examples/embedding-provider-rs/README.md

@@ -98,7 +98,7 @@ pip install cmake
 
 ## Differences from Python Provider
 
-| 维度 | Python (`tools/embedding-provider/`) | Rust (`tools/embedding-provider-rs/`) |
+| 维度 | Python (`examples/embedding-provider/`) | Rust (`examples/embedding-provider-rs/`) |
 |------|--------------------------------------|---------------------------------------|
 | 依赖 | `requests` (HTTP) → Ollama | `embellama` (本地 GGUF 直接加载) |
 | 运行时 | 需要 Ollama 服务 | 零外部服务，纯本地推理 |

tools/embedding-provider/README.md → examples/embedding-provider/README.md

@@ -26,7 +26,7 @@ devbase follows an **"outboard brain"** design: the Rust core handles storage, i
 
 ```bash
 # 1. Clone or navigate into this directory
-cd tools/embedding-provider
+cd examples/embedding-provider
 
 # 2. Install dependencies
 pip install -r requirements.txt

tools/invariant-checks/run-checks.ps1 → scripts/invariant-checks/run-checks.ps1

@@ -2,7 +2,7 @@
 # SPDX-License-Identifier: MIT
 # Copyright (c) 2026 juice094
 # devbase Architecture Invariant CI Checks
-# Run from repo root: tools/invariant-checks/run-checks.ps1
+# Run from repo root: scripts/invariant-checks/run-checks.ps1
 
 $ErrorActionPreference = "Stop"
 $script:Failed = 0

skills/embed-repo/SKILL.md

@@ -21,7 +21,7 @@ outputs:
 ---
 # Embed Repository
 
-This skill invokes the local embedding provider (`tools/embedding-provider/local.py`)
+This skill invokes the local embedding provider (`examples/embedding-provider/local.py`)
 to generate 384-dimensional vectors for all code symbols in the specified repository.
 
 ## Usage