memorix 1.0.1 → 1.0.3

package/CHANGELOG.md

@@ -2,6 +2,33 @@
 
 All notable changes to this project will be documented in this file.
 
+## [1.0.3] — 2026-03-14
+
+### Added
+- **Memory Formation Pipeline** — Three-stage pipeline (Extract → Resolve → Evaluate) runs in shadow mode on every `memorix_store` call and hooks trigger. Collects quality metrics without affecting storage decisions.
+  - **Extract**: Automatic fact extraction from narratives, title normalization, entity resolution against Knowledge Graph, observation type verification.
+  - **Resolve**: 4 resolution actions (new/merge/evolve/discard) based on similarity scoring, word overlap, and contradiction detection.
+  - **Evaluate**: Multi-factor knowledge value assessment (type weight, fact density, specificity, causal reasoning, noise detection). Categorizes memories as core/contextual/ephemeral.
+- **`memorix_formation_metrics` tool** — New MCP tool to query aggregated Formation Pipeline metrics (value scores, resolution actions, extraction rates, processing times).
+- **`getEntityNames()` method** on `KnowledgeGraphManager` for Formation Pipeline entity resolution.
+
+### Stats
+- **Default MCP Tools:** 23 (+1: `memorix_formation_metrics`)
+- **Tests:** 803/803 passing across 60 files (+50 new Formation Pipeline tests)
+- **Hooks safety:** handler.ts +21 lines (shadow call only), zero modification to existing hook logic
+
+---
+
+## [1.0.2] — 2026-03-14
+
+### Fixed
+- **MCP Server version mismatch** — Server now reports the correct version from `package.json` (was hardcoded `0.1.0`). Injected at build time via tsup `define`.
+- **CI Node.js matrix** — Removed Node 18 from CI matrix to match `engines: >=20` in `package.json`.
+- **Orama reindex idempotency** — `reindexObservations()` now resets the Orama DB before rebuilding, eliminating "document already exists" errors in multi-session scenarios.
+- **E2E tests no longer touch real user data** — Mini-skills E2E tests now use a temporary directory with synthetic observations instead of reading/writing `~/.memorix/data/`.
+
+---
+
 ## [1.0.1] — 2026-03-14
 
 ### Fixed

package/README.md

@@ -14,12 +14,12 @@
   <a href="https://www.npmjs.com/package/memorix"><img src="https://img.shields.io/npm/dm/memorix.svg?style=flat-square&color=blue" alt="downloads"></a>
   <a href="LICENSE"><img src="https://img.shields.io/badge/license-Apache%202.0-green.svg?style=flat-square" alt="license"></a>
   <a href="https://github.com/AVIDS2/memorix"><img src="https://img.shields.io/github/stars/AVIDS2/memorix?style=flat-square&color=yellow" alt="stars"></a>
-  <img src="https://img.shields.io/badge/tests-753%20passed-brightgreen?style=flat-square" alt="tests">
+  <img src="https://img.shields.io/badge/tests-803%20passed-brightgreen?style=flat-square" alt="tests">
   <a href="https://github.com/AVIDS2/memorix/actions/workflows/ci.yml"><img src="https://img.shields.io/github/actions/workflow/status/AVIDS2/memorix/ci.yml?style=flat-square&label=CI" alt="CI"></a>
 </p>
 
 <p align="center">
-  <strong>v1.0 Stable | 22 MCP tools | Auto-cleanup | Multi-agent collaboration | 10 IDEs supported</strong>
+  <strong>v1.0 Stable | 23 MCP tools | Auto-cleanup | Multi-agent collaboration | 10 IDEs supported</strong>
 </p>
 
 <p align="center">
@@ -64,6 +64,7 @@ No re-explaining. No copy-pasting. No vendor lock-in.
 - **Dual-Mode Quality**: Free heuristic engine for basic dedup; optional LLM mode for intelligent compression, reranking, and conflict resolution.
 - **3-Layer Progressive Disclosure**: Search returns compact indices (~50 tokens/result), timeline shows chronological context, detail provides full content. ~10x token savings over full-text retrieval.
 - **Mini-Skills**: Promote high-value observations to permanent skills that auto-inject at every session start. Critical knowledge never decays.
+- **Memory Formation Pipeline**: Automatic fact extraction, entity resolution, and knowledge value assessment on every store. Shadow mode collects quality metrics without affecting storage.
 - **Auto-Memory Hooks**: Automatically capture decisions, errors, and gotchas from IDE tool calls. Pattern detection in English and Chinese.
 - **Knowledge Graph**: Entity-relation model compatible with [MCP Official Memory Server](https://github.com/modelcontextprotocol/servers/tree/main/src/memory). Auto-creates relations from entity extraction.
 
@@ -171,7 +172,7 @@ Restart your agent. No API keys required. No cloud. No external dependencies.
 
 ## Features
 
-### 22 MCP Tools (Default)
+### 23 MCP Tools (Default)
 
 | Category | Tools |
 |----------|-------|
@@ -179,7 +180,7 @@ Restart your agent. No API keys required. No cloud. No external dependencies.
 | **Sessions** | `memorix_session_start` · `memorix_session_end` · `memorix_session_context` |
 | **Skills** | `memorix_skills` · `memorix_promote` |
 | **Workspace** | `memorix_workspace_sync` · `memorix_rules_sync` |
-| **Maintenance** | `memorix_retention` · `memorix_consolidate` · `memorix_transfer` |
+| **Maintenance** | `memorix_retention` · `memorix_consolidate` · `memorix_transfer` · `memorix_formation_metrics` |
 | **Team** | `team_manage` · `team_file_lock` · `team_task` · `team_message` |
 | **Dashboard** | `memorix_dashboard` |
 
@@ -329,7 +330,7 @@ Stage 3:  Recency + Project Affinity    →  Final scored results
 ### Write Pipeline
 
 ```
-Input  →  LLM Compression (optional)  →  Compact on Write (dedup/merge)  →  Store + Index
+Input  →  Formation Pipeline (extract/resolve/evaluate)  →  LLM Compression (optional)  →  Compact on Write (dedup/merge)  →  Store + Index
 ```
 
 ### Key Design Decisions
@@ -348,7 +349,7 @@ git clone https://github.com/AVIDS2/memorix.git
 cd memorix && npm install
 
 npm run dev       # watch mode
-npm test          # 753 tests
+npm test          # 803 tests
 npm run build     # production build
 ```
 

package/README.zh-CN.md

@@ -14,12 +14,12 @@
   <a href="https://www.npmjs.com/package/memorix"><img src="https://img.shields.io/npm/dm/memorix.svg?style=flat-square&color=blue" alt="downloads"></a>
   <a href="LICENSE"><img src="https://img.shields.io/badge/license-Apache%202.0-green.svg?style=flat-square" alt="license"></a>
   <a href="https://github.com/AVIDS2/memorix"><img src="https://img.shields.io/github/stars/AVIDS2/memorix?style=flat-square&color=yellow" alt="stars"></a>
-  <img src="https://img.shields.io/badge/tests-753%20passed-brightgreen?style=flat-square" alt="tests">
+  <img src="https://img.shields.io/badge/tests-803%20passed-brightgreen?style=flat-square" alt="tests">
   <a href="https://github.com/AVIDS2/memorix/actions/workflows/ci.yml"><img src="https://img.shields.io/github/actions/workflow/status/AVIDS2/memorix/ci.yml?style=flat-square&label=CI" alt="CI"></a>
 </p>
 
 <p align="center">
-  <strong>v1.0 正式版 | 22 个 MCP 工具 | 自动清理 | 多 Agent 协作 | 10 IDE 支持</strong>
+  <strong>v1.0 正式版 | 23 个 MCP 工具 | 自动清理 | 多 Agent 协作 | 10 IDE 支持</strong>
 </p>
 
 <p align="center">
@@ -64,6 +64,7 @@ AI 编码 Agent 在会话之间会丢失全部上下文。切换 IDE 后，之
 - **双模式质量引擎**：免费启发式引擎处理基础去重；可选 LLM 模式提供智能压缩、重排序和冲突检测。
 - **3 层渐进式展示**：搜索返回紧凑索引（每条约 50 tokens），时间线展示前后文，详情提供完整内容。相比全文检索节省约 10 倍 token。
 - **Mini-Skills**：将高价值观察提升为永久技能，每次会话启动自动注入。关键知识永不衰减。
+- **记忆形成管线**：每次存储自动进行事实抽取、实体归并和知识价值评估。Shadow 模式收集质量指标而不影响存储。
 - **自动记忆 Hook**：自动从 IDE 工具调用中捕获决策、错误和踩坑经验。支持中英文模式检测。
 - **知识图谱**：实体-关系模型，兼容 [MCP 官方 Memory Server](https://github.com/modelcontextprotocol/servers/tree/main/src/memory)。自动从内容中提取实体并创建关联。
 
@@ -171,7 +172,7 @@ args = ["serve"]
 
 ## 功能
 
-### 22 个 MCP 工具（默认）
+### 23 个 MCP 工具（默认）
 
 | 分类 | 工具 |
 |------|------|
@@ -179,7 +180,7 @@ args = ["serve"]
 | **会话** | `memorix_session_start` · `memorix_session_end` · `memorix_session_context` |
 | **技能** | `memorix_skills` · `memorix_promote` |
 | **工作区** | `memorix_workspace_sync` · `memorix_rules_sync` |
-| **维护** | `memorix_retention` · `memorix_consolidate` · `memorix_transfer` |
+| **维护** | `memorix_retention` · `memorix_consolidate` · `memorix_transfer` · `memorix_formation_metrics` |
 | **团队** | `team_manage` · `team_file_lock` · `team_task` · `team_message` |
 | **仪表盘** | `memorix_dashboard` |
 
@@ -307,7 +308,7 @@ memorix hooks install # 为 IDE 安装自动记忆
 graph TB
     A["Cursor · Claude Code · Windsurf · Codex · +6 更多"]
     A -->|MCP stdio| Core
-    Core["Memorix MCP Server\n22 个默认工具 · 自动Hook · 自动清理"]
+    Core["Memorix MCP Server\n23 个默认工具 · 自动Hook · 自动清理"]
     Core --> Search["检索管线\nBM25 + 向量 + 重排序"]
     Core --> Team["团队协作\n注册 · 任务 · 锁 · 消息"]
     Core --> Sync["规则 & 工作区同步\n10 个适配器"]
@@ -331,7 +332,7 @@ graph TB
 ### 写入管线
 
 ```
-输入  →  LLM 压缩（可选）  →  写入时去重/合并  →  存储 + 索引
+输入  →  记忆形成管线 (抽取/归并/评估)  →  LLM 压缩（可选）  →  写入时去重/合并  →  存储 + 索引
 ```
 
 ### 关键设计决策
@@ -350,7 +351,7 @@ git clone https://github.com/AVIDS2/memorix.git
 cd memorix && npm install
 
 npm run dev       # 监听模式
-npm test          # 753 个测试
+npm test          # 803 个测试
 npm run build     # 生产构建
 ```