memorix 1.0.9 → 1.0.10

package/CHANGELOG.md

@@ -2,6 +2,25 @@
 
 All notable changes to this project will be documented in this file.
 
+## [1.0.10] - 2026-06-08
+
+### Added -- Low-Intrusion Handoff Diagnostics
+- **Privacy-safe handoff receipts (#95)** -- Added `memorix receipt --json` for cross-agent memory debugging. Receipts expose project identity hashes, write/search counts, recent observation ID hashes, and privacy boundaries without raw chat, raw memory text, raw search queries, tool payloads, or local file paths.
+- **Doctor receipt mode** -- Added `memorix doctor --receipt` so support/debugging can include the same receipt summary alongside normal health diagnostics.
+- **Cross-agent memory boundary docs** -- Documented the operational boundary: MCP connection success does not mean a memory was written, and shared memory means stored memories are searchable across clients in the same Git project, not that chat transcripts are mirrored.
+- **TUI Agent LLM config scope** -- Added `agent` / `MEMORIX_AGENT_LLM_*` configuration for TUI/chat agent calls. Memory formation, compression, rerank, and embedding remain on their existing `llm` / `embedding` configuration paths. If `agent` is unset, TUI chat falls back to `llm` for backward compatibility.
+
+### Changed
+- **Generated agent rules are less intrusive** -- Rules now prefer `memorix_search` when useful and treat `memorix_session_start` as optional, reserved for handoff, long-running work, session recovery, team coordination, or HTTP project binding.
+- **Agent Team wording tightened** -- Public docs now frame Agent Team as optional autonomous-agent/subagent coordination state rather than an IDE-window chat room. Existing `team_*` APIs and CLI namespaces are unchanged.
+
+### Fixed
+- **Release publish blocker** -- `tests/integration/release-blockers.test.ts` now reserves an ephemeral local TCP port for the real `memorix serve-http` smoke test instead of assuming port `19879` is free. This removes a CI-only `EADDRINUSE` failure that blocked the npm publish workflow.
+- **Standalone dashboard embedding config (#46 follow-up)** -- `memorix dashboard` now loads project `.env` / project YAML before dashboard status routes initialize embedding provider state, matching `status`, `doctor`, and TUI behavior.
+
+### Documentation
+- Added a small moving-between-machines note covering `memorix transfer export/import` without changing the Memorix package name or primary README structure.
+
 ## [1.0.9] - 2026-05-19
 
 ### Added -- Knowledge Base / LLM Wiki Foundation

package/README.md

@@ -52,7 +52,7 @@ Most coding agents remember only the current thread. Memorix gives them a shared
 <tr><td><b>🔍 Source-Aware Retrieval</b></td><td>"What changed" queries favor Git Memory; "why" queries favor reasoning; project-scoped by default, global on demand</td></tr>
 <tr><td><b>⚙️ Memory Quality Pipeline</b></td><td>Formation (LLM-assisted evaluation), dedup, consolidation, retention with exponential decay — memory stays clean, not noisy</td></tr>
 <tr><td><b>🔄 Workspace & Rules Sync</b></td><td>One command to migrate MCP configs, workflows, rules, and skills across Cursor, Windsurf, Claude Code, Codex, Copilot, Kiro, etc.</td></tr>
-<tr><td><b>👥 Agent Team</b></td><td>Opt-in autonomous-agent state: task board with role-based claiming, inter-agent messaging, advisory file locks, situational-awareness poll</td></tr>
+<tr><td><b>👥 Agent Team</b></td><td>Opt-in autonomous-agent state: task board with role-based claiming, handoff messages, advisory file locks, situational-awareness poll</td></tr>
 <tr><td><b>🤖 Multi-Agent Orchestration</b></td><td><code>memorix orchestrate</code> runs a structured coordination loop — plan → parallel execution → verify → fix → review — with capability routing and worktree isolation</td></tr>
 <tr><td><b>📋 Session Lifecycle</b></td><td>Session start/end with handoff summaries, watermark tracking (new memories since last session), cross-session context recovery</td></tr>
 <tr><td><b>🎯 Project Skills</b></td><td>Auto-generate SKILL.md from memory patterns; promote observations to permanent mini-skills injected at session start</td></tr>
@@ -241,9 +241,10 @@ For the full IDE matrix, Windows notes, and troubleshooting, see [docs/SETUP.md]
 | Capture Git truth | `memorix git-hook --force`, `memorix ingest commit`, `memorix ingest log` | [Git Memory Guide](docs/GIT_MEMORY.md) |
 | Run dashboard + HTTP MCP | `memorix background start` | [Setup Guide](docs/SETUP.md), [Docker](docs/DOCKER.md) |
 | Keep memory-only sessions lightweight | `memorix_session_start(projectRoot=...)` or `memorix session start` | [Agent Operator Playbook](docs/AGENT_OPERATOR_PLAYBOOK.md#8-what-an-agent-should-do-at-session-start) |
-| Join the autonomous agent team | `memorix session start --joinTeam` or `memorix team join` | [TEAM.md](TEAM.md), [API Reference](docs/API_REFERENCE.md#9-agent-team-tools) |
-| Run autonomous multi-agent work | `memorix orchestrate --goal "..."` | [API Reference](docs/API_REFERENCE.md) |
-| Sync agent configs/rules | `memorix sync workspace ...`, `memorix sync rules ...` | [Setup Guide](docs/SETUP.md) |
+| Join the autonomous agent team | `memorix session start --joinTeam` or `memorix team join` | [TEAM.md](TEAM.md), [API Reference](docs/API_REFERENCE.md#9-agent-team-tools) |
+| Run autonomous multi-agent work | `memorix orchestrate --goal "..."` | [API Reference](docs/API_REFERENCE.md) |
+| Debug cross-agent memory handoff | `memorix receipt --json` or `memorix doctor --receipt` | Privacy-safe hashes/counts only; no raw chat or memory text |
+| Sync agent configs/rules | `memorix sync workspace ...`, `memorix sync rules ...` | [Setup Guide](docs/SETUP.md) |
 | Use Memorix from code | `import { createMemoryClient } from 'memorix/sdk'` | [API Reference](docs/API_REFERENCE.md) |
 
 The most common loop is deliberately small:
@@ -262,7 +263,9 @@ HTTP MCP sessions idle out after 30 minutes by default. If your client does not
 MEMORIX_SESSION_TIMEOUT_MS=86400000 memorix background start  # 24h
 ```
 
-Agent Team is **not** the normal memory startup path and it is **not** a chat room between IDE windows. Join only when you need tasks, messages, locks, or a structured autonomous-agent workflow. For real multi-agent execution, prefer:
+Shared memory means stored memories are searchable across clients bound to the same Git project. It does not mean every chat message is mirrored automatically. For support/debugging, `memorix receipt --json` and `memorix doctor --receipt` show project identity, write/search counts, and hashed IDs without raw prompts or memory text.
+
+Agent Team is **not** the normal memory startup path and it is **not** a chat room between IDE windows. Join only when you need tasks, handoff messages, locks, or a structured autonomous-agent/subagent workflow. For real multi-agent execution, prefer:
 
 ```bash
 memorix orchestrate --goal "Add user authentication" --agents claude-code,cursor,codex
@@ -431,17 +434,16 @@ Additional deep references:
 
 ---
 
-## What's New in 1.0.9
+## What's New in 1.0.10
 
-Version `1.0.9` adds the first readable project knowledge layer and turns the terminal UI into a knowledge-native workbench.
+Version `1.0.10` tightens Memorix's low-intrusion operator story: privacy-safe handoff diagnostics, clearer cross-agent memory boundaries, optional session semantics in generated rules, independent TUI agent LLM config, a better fresh-project cold-start path, safer update behavior, and a release-gate fix so the real `serve-http` smoke test no longer depends on port `19879` being free on the runner.
 
-- **Knowledge Base / LLM Wiki**: Memorix now exposes a generated, readable Knowledge Base from durable observations, git facts, mini-skills, and project evidence. Raw memory remains the source of truth; the Knowledge Base is a citable synthesis for humans and agents.
-- **Semantic Knowledge Graph**: A graph projection over the same eligible knowledge inputs preserves source refs and supports concept/module/decision exploration without claiming to be GraphRAG.
-- **Tabbed TUI Workbench**: `Home`, `Knowledge`, `Memory`, `Workbench`, and `Graph` tabs replace the old scattered view model.
-- **Cross-Reference Navigation**: Jump between Knowledge items, Memory detail, and Graph nodes through stable refs where available.
-- **Explicit Session Center**: Workbench shows session state and context sources with explicit bind/end actions. It does not auto-start sessions or auto-join Agent Team.
-- **TUI Chat Usability Fixes**: Chat input remains active after assistant responses, and view shortcuts no longer steal normal CommandBar text.
-- **Architecture Boundaries**: App, realtime agent messaging, full coding-agent harness, graph editing, and GraphRAG remain outside 1.0.9 scope.
+- **Privacy-Safe Handoff Receipt**: `memorix receipt --json` and `memorix doctor --receipt` expose hashes/counts for project identity, writes, and optional probe results without leaking raw chat, memory text, queries, tool payloads, or local paths.
+- **Lower-Intrusion Agent Guidance**: Generated rules now treat `memorix_session_start` as optional and avoid repeating `memorix_search` in the same turn after a confirmed fresh-project cold start.
+- **Separate TUI Agent LLM Config**: `agent` / `MEMORIX_AGENT_LLM_*` lets TUI chat use a different provider/model from memory formation and rerank, while preserving fallback to `llm`.
+- **Fresh-Project UX Fix**: Empty project searches now explicitly say the tool call worked but the project has no stored Memorix memories yet, instead of looking like a generic retrieval failure.
+- **Safer Update Checks**: Auto-update now defaults to notify-only; background install is explicit opt-in via `MEMORIX_AUTO_UPDATE=install`.
+- **Dashboard Config Consistency**: Standalone dashboard status now loads project `.env` and project YAML before embedding/LLM status routes initialize, matching CLI and TUI behavior.
 
 ---
 

package/README.zh-CN.md

@@ -431,17 +431,16 @@ Memorix 不是一条单线流水线。它从多个入口接收记忆，把内容
 
 ---
 
-## 1.0.9 更新亮点
+## 1.0.10 更新亮点
 
-`1.0.9` 增加了第一版可读项目知识层，并把终端 UI 收束为 knowledge-native workbench。
+`1.0.10` 继续把 Memorix 往更低侵入、更容易排障的 operator 体验上收紧：跨 agent 交接诊断更安全、共享记忆边界更清晰、生成规则更克制、TUI agent LLM 配置更独立，同时补上了一个发版阻断修复：真实 `serve-http` 冒烟测试改为运行时申请空闲本地端口，不再依赖 CI 机器上的 `19879` 端口必须空闲。
 
-- **Knowledge Base / LLM Wiki**：Memorix 现在可以从 durable observations、Git facts、mini-skills 和项目证据生成可读的 Knowledge Base。Raw memory 仍是事实来源；Knowledge Base 是给人和 agent 使用的可引用综合层。
-- **Semantic Knowledge Graph**：基于同一批 eligible knowledge inputs 生成语义图谱投影，保留 source refs，支持围绕概念、模块和决策探索，但不把它包装成 GraphRAG。
-- **Tabbed TUI Workbench**：`Home`、`Knowledge`、`Memory`、`Workbench`、`Graph` 五个 tab 替代旧的分散视图模型。
-- **跨引用导航**：可通过稳定 refs 在 Knowledge item、Memory detail 和 Graph node 之间跳转。
-- **显式 Session Center**：Workbench 展示 session 状态和上下文来源，bind/end 都是显式动作；进入 tab 不会自动 start session，也不会自动 join Agent Team。
-- **TUI Chat 可用性修复**：assistant 回复后输入框保持可用；Graph/Memory/Knowledge 的视图快捷键不再吞掉正常 CommandBar 输入。
-- **架构边界**：desktop app、realtime agent messaging、完整 coding-agent harness、graph editing 和 GraphRAG 都不属于 1.0.9 范围。
+- **隐私安全的交接 Receipt**：`memorix receipt --json` 和 `memorix doctor --receipt` 提供项目身份哈希、写入/检索计数、最近 observation ID 哈希和可选 probe 结果计数，用于跨 agent memory 排障；不暴露原始 chat、记忆文本、查询、tool payload 或本地路径。
+- **更低侵入的 Agent 指南**：生成规则现在把 `memorix_session_start` 视为可选动作，只在 handoff、长任务、session 恢复、team 协调或 HTTP 项目绑定时推荐使用。
+- **独立的 TUI Agent LLM 配置**：新增 `agent` / `MEMORIX_AGENT_LLM_*` 配置路径，允许 TUI chat 使用与 memory formation、rerank、embedding 不同的 provider/model；未配置时仍回退到 `llm`。
+- **Fresh Project 体验修复**：空项目搜索现在会明确说明「工具调用成功，但当前 Git 项目还没有 Memorix memories」，而不是表现得像一个泛化检索失败。
+- **更安全的更新检查**：自动更新默认切回 notify-only；后台静默安装需要显式设置 `MEMORIX_AUTO_UPDATE=install`。
+- **Dashboard 配置一致性**：独立 `memorix dashboard` 现在会在状态路由初始化 embedding/LLM 前先加载项目 `.env` 和 YAML，和 CLI、TUI、doctor/status 保持一致。
 
 ---