lumina-wiki 1.5.0 → 1.6.0

package/CHANGELOG.md

@@ -3,6 +3,90 @@
 All notable changes to Lumina-Wiki are documented here.
 Format follows [Keep a Changelog](https://keepachangelog.com/en/1.0.0/).
 
+## [Unreleased]
+
+## [1.6.0] - 2026-05-15
+
+### Added — Multi-provider PDF resolution + RSS / Atom feeds (research pack)
+
+- **OpenAlex fetcher** activated as the metadata anchor across the new ladder
+  (Phase 1–2). `external_ids.openalex` namespace now persists Work IDs
+  (`^W\d+$`).
+- **`fetch_unpaywall.py`** — DOI → best OA PDF URL. Requires
+  `UNPAYWALL_EMAIL` (free email-of-record).
+- **`fetch_core.py`** — CORE search + download-url. Optional
+  `CORE_API_KEY`; ladder skips CORE on 429 and warns once.
+- **`resolve_pdf.py`** — 2-layer orchestrator. Layer A always runs OpenAlex
+  (cross-walks DOI ↔ arXiv ↔ OpenAlex). Layer B is a stop-on-first-200 PDF
+  ladder: `oa_url → unpaywall → core → arxiv`. Each provider attempt is
+  logged to stderr; the final shape carries `external_ids`, `sources[]`,
+  `pdf_path`, and `status` (`ok` | `metadata_only` | `failed`).
+- **`fetch_rss.py`** — RSS / Atom poller with etag caching, defusedxml-based
+  XXE rejection, per-feed state files under `_lumina/_state/feeds/<id>.json`,
+  spill-aware `max_new` cap, and 5000-entry / 90-day `last_seen_guids`
+  eviction.
+- **Watchlist `type: feed`** items extend `_lumina/config/watchlist.yml`
+  additively. v1 files without `type` keep validating (defaults to
+  `topic`). Feed URLs are gated to `^https://` and rejected if they start
+  with `--` (flag-injection defense-in-depth).
+- **`/lumi-research-watch-run`** skill orchestrates a single pass over the
+  consolidated watchlist (topics + feeds). User owns scheduling — three
+  patterns documented (cron, launchd, Task Scheduler).
+- **`cron-daily.sh` wrapper** ships under
+  `_lumina/scripts/scheduler-samples/`. Inert until the user wires it into
+  their scheduler. Sets `umask 077`, `chmod 600` on the log,
+  rotates at 1 MB.
+- **`extract_ids_from_text()`** in `id_utils.py` — reusable free-text
+  identifier harvester for feed entry titles / summaries / link hrefs.
+
+### Added — Project governance
+
+- `CONTRIBUTING.md` at the repo root: workflow checklists for adding skills,
+  fetchers, schema changes, installer changes, and entry-point stubs; the
+  trilingual user-docs convention; CI gates; exit-code contract; and a section
+  specifically scoped to AI-agent contributors that points at
+  `docs/project-context.md`, `CLAUDE.md`, and `docs/DEVELOPMENT.md` as
+  load-bearing context.
+- `CODE_OF_CONDUCT.md` at the repo root: Contributor Covenant v2.1, contact
+  `tronghieu.luu@gmail.com`. Linked from `CONTRIBUTING.md` §2.
+- `SECURITY.md` at the repo root: supported-versions table, private
+  reporting channels (GitHub Private Vulnerability Reporting + email),
+  in-scope / out-of-scope surfaces, severity bands, and coordinated
+  disclosure expectations.
+- `.github/PULL_REQUEST_TEMPLATE.md`: per-change-type checklists that
+  mirror `CONTRIBUTING.md` §5 (skill / fetcher / schema / installer),
+  trilingual docs checkpoint, and a rule-deviation prompt.
+
+### Security
+
+- **SSRF guard** (`_safe_url`) on every PDF candidate URL: rejects RFC1918,
+  loopback, link-local, multicast, cloud-metadata (169.254.169.254).
+  Re-validated post-redirect.
+- **`fetch_pdf.py` mid-stream size cap** (`MAX_PDF_BYTES = 100 MiB`) — a
+  malicious endpoint that lies about Content-Length now aborts mid-download
+  and cleans up `.tmp`.
+- **DOI filename hashing** — DOIs are hashed to a 16-char SHA-256 prefix on
+  disk to neutralize Windows-reserved-name collisions (CON, PRN, AUX, NUL).
+- **XXE pre-parse** — every RSS / Atom body is run through
+  `defusedxml.ElementTree.fromstring` before feedparser sees it; DOCTYPE /
+  billion-laughs payloads are rejected without state mutation.
+
+### Requirements
+
+- New optional env vars: `UNPAYWALL_EMAIL`, `CORE_API_KEY`. Both gracefully
+  skip (ladder continues) when unset.
+- `requirements.txt` adds `feedparser>=6.0` (research pack only).
+
+### Backwards compatibility
+
+- `external_ids.openalex` is additive — existing pages continue to validate.
+- `sources[]` is additive — entries without an entry stay valid.
+  `ns/value` fields drop silently if either is missing or invalid (same
+  forgiveness model as the existing `url` field).
+- Watchlist v1 (no `type` field) still validates and runs unchanged.
+- `fetch_pdf.py` CLI is stable; new helpers (`_safe_url`, `head_check`,
+  `MAX_PDF_BYTES`) are additions only.
+
 ## [1.5.0] - 2026-05-10
 
 ### Added — Learning Pack: `/lumi-learning-reflect` self-reflection skill (PRs #16, #17)
@@ -398,7 +482,8 @@ Format follows [Keep a Changelog](https://keepachangelog.com/en/1.0.0/).
 
 ---
 
-[Unreleased]: https://github.com/tronghieu/lumina-wiki/compare/v1.4.0...HEAD
+[Unreleased]: https://github.com/tronghieu/lumina-wiki/compare/v1.5.0...HEAD
+[1.5.0]: https://github.com/tronghieu/lumina-wiki/compare/v1.4.0...v1.5.0
 [1.4.0]: https://github.com/tronghieu/lumina-wiki/compare/v1.3.0...v1.4.0
 [1.3.0]: https://github.com/tronghieu/lumina-wiki/compare/v1.2.0...v1.3.0
 [1.2.0]: https://github.com/tronghieu/lumina-wiki/compare/v1.1.0...v1.2.0

package/README.md

@@ -184,6 +184,7 @@ These are the commands you can use when chatting with your AI agent.
 | | `/lumi-research-prefill` | Seed foundational concepts to avoid duplicates. |
 | | `/lumi-research-topic` | Create a topic page at `wiki/topics/<slug>.md` by gathering related concepts and sources already in your wiki. The AI proposes what to include and you confirm before anything is written. Use this after several `/lumi-ingest` runs when you want to give a theme its own page. |
 | | `/lumi-research-setup` | Help configure API keys for research tools. |
+| | `/lumi-research-watch-run` | Run one scheduled-discovery pass over your watchlist (topics + RSS / Atom feeds). Polls only when you ask. |
 | **Reading** | `/lumi-reading-chapter-ingest` | Ingest a book chapter by chapter. |
 | | `/lumi-reading-character-track` | Track characters and their relationships in a story. |
 | | `/lumi-reading-theme-map` | Identify and map themes in a narrative. |
@@ -204,8 +205,8 @@ Lumina-Wiki is evolving rapidly. Here is our user-facing roadmap:
 - [x] **Multilingual setup:** Choose English, Vietnamese, or Chinese as your primary language during install. *(shipped in v1.2)*
 - [x] **Native DOCX, RTF & EPUB ingestion:** Pull Word, Rich Text, and EPUB books straight into your wiki via the research pack. *(shipped in v1.x)*
 - [x] **Improved CI/CD:** Native support for Bun and Node 22 environments. *(shipped in v1.2)*
-- [ ] **Global Source Expansion:** Direct integration with OpenAlex, CORE, and Unpaywall for reliable DOI-to-PDF resolution.
-- [ ] **RSS & Blog Monitoring:** Automatically surface new papers from your favorite lab blogs and journals.
+- [x] **Global Source Expansion:** Direct integration with OpenAlex, CORE, and Unpaywall for reliable DOI-to-PDF resolution. *(shipped in v1.6)*
+- [x] **RSS & Blog Monitoring:** Automatically surface new papers from your favorite lab blogs and journals via `type: feed` watchlist items. *(shipped in v1.6)*
 - [ ] **Advanced Paper Ranking:** See influence scores and quality signals for your research papers.
 
 **Long-term (Deep Research & Integration)**
@@ -250,3 +251,198 @@ npm run test:all
 - [简体中文 (Chinese)](README.zh.md)
 
 **License:** [MIT](LICENSE) © Lưu Trọng Hiếu.
+
+<!-- lumina:schema -->
+
+## Roles
+
+You are the wiki maintainer. The user curates sources, asks questions, and directs analysis. You do everything else: read, summarize, connect pages, file notes, run health checks, and keep the wiki coherent. You write the wiki; the user reads it.
+
+Always communicate with the user in **English**. Always write wiki pages in **English**.
+
+### User Communication
+
+- Default to a clear, everyday style suitable for most users. You are a helpful knowledge assistant, not a software engineer explaining implementation details.
+- Use **English** for every conversational message. Do not mix languages unless quoting source text, file names, commands, or proper nouns.
+- Translate workflow terms into the user's language. If a source uses an important domain term, write the translated term first and put the original term in parentheses on first use.
+- Speak to non-technical users. Use short, natural sentences. Say what the user gets, what changed, what needs attention, or what decision is needed; keep internal tool details quiet unless the user asks.
+- Prefer plain phrases such as "checking links", "checking against the source", "saving the page", and "I found something to review" over tool-centric words like lint, schema, frontmatter, checkpoint, verify, or JSON in user-facing messages.
+- If technical detail is necessary, give the plain-language meaning first, then the technical term in parentheses.
+- Ask the user only when their judgment is needed: approving a draft, choosing between ambiguous sources, allowing an overwrite/restart, handling source-check findings, accepting lower confidence, or deciding how to fix an issue the tools cannot fix safely.
+
+---
+
+## Repository Layout
+
+Keep this mental map in immediate context:
+
+### `wiki/` is the main product surface
+
+- `wiki/index.md` — catalog of all wiki pages, updated on every ingest
+- `wiki/log.md` — append-only activity log
+- `wiki/concepts/` — reusable knowledge structure
+- `wiki/sources/` — per-source summaries (papers, articles, books, podcasts, notes)
+- `wiki/people/` — people referenced across sources
+- `wiki/summary/` — area-level syntheses
+- `wiki/outputs/` — generated artifacts (comparisons, exports)
+- `wiki/graph/` — derived state; never edit manually
+
+### `raw/` is user-owned
+
+- `raw/sources/` — `.pdf`, `.tex`, `.html`, `.md`, transcripts, anything ingested
+- `raw/notes/` — user's own markdown notes
+- `raw/assets/` — images and binary attachments
+- `raw/tmp/` — sidecar files generated by skills (transient; do not store canonical sources here)
+- `raw/download/<resource>/` — full-text artifacts auto-fetched by skills, partitioned by source
+  (e.g. `raw/download/arxiv/2604.03501v2.pdf`, `raw/download/doi/<doi>.pdf`).
+  Permanent agent-writable zone — keep separate from `raw/sources/` (human-curated).
+
+**Rule:** never modify or delete an existing file under `raw/`. Files added by the user are authoritative and immutable to the agent. New files may only be *added*, only by a skill that documents this behavior, and only into `raw/tmp/`, `raw/download/`. Every other path under `raw/` is read-only.
+
+### `.agents/` is the skill source of truth
+
+- `.agents/skills/lumi-*/` — installed skills (flat, one directory per skill)
+
+### `_lumina/` is the installer-managed sidecar
+
+- `_lumina/config/lumina.config.yaml` — workspace config; editable
+- `_lumina/schema/` — deeper reference docs; open when this file points you there
+- `_lumina/scripts/` — Node engine (`wiki.mjs`, `lint.mjs`, `reset.mjs`, `schemas.mjs`)
+- `_lumina/tools/` — Python tools (always: `extract_pdf.py`, `fetch_pdf.py`, `requirements.txt`)
+- `_lumina/_state/` — installer/skill checkpoint state; gitignored
+- `_lumina/manifest.json` — installer state; never edit by hand
+
+---
+
+## Page Types
+
+Every wiki page has a defined type, frontmatter, and section structure. **Open `_lumina/schema/page-templates.md` before drafting a new page or repairing an existing one** — it has the full templates and required frontmatter fields.
+
+| Type       | Directory      | Purpose                                                              |
+|------------|---------------|----------------------------------------------------------------------|
+| Source     | `sources/`    | Per-document summary: key claims, evidence, takeaways, questions     |
+| Concept    | `concepts/`   | Cross-source idea or technique with variants and comparisons         |
+| Person     | `people/`     | Profile of a referenced person with key sources and relationships    |
+| Summary    | `summary/`    | Area-level synthesis spanning multiple sources and concepts          |
+
+---
+
+## Link Syntax
+
+All internal links use Obsidian wikilinks:
+
+```markdown
+[[slug]]                     — link to any page in this wiki
+[[chain-of-thought]]         — links to concepts/chain-of-thought.md
+[[1984-orwell]]              — links to sources/1984-orwell.md
+```
+
+**Slug rule**: lowercase, hyphen-separated, no spaces, no diacritics.
+
+---
+
+## Cross-Reference Rules (Bidirectional Links)
+
+When you write a forward link, **always write the reverse link in the same operation**. This is the heart of why the wiki compounds. Skipping it leaves the graph half-built.
+
+| Forward action                              | Required reverse action                    |
+|---------------------------------------------|---------------------------------------------|
+| `sources/A` writes `Related: [[concept-B]]` | `concepts/B` appends A to `Key sources`    |
+| `sources/A` writes `[[person-C]]`           | `people/C` appends A to `Key sources`      |
+| `concepts/K` writes `[[source-E]]`          | `sources/E` appends K to `Related concepts`|
+| `summary/S` writes `[[concept-K]]`          | `concepts/K` appends S to `Mentioned in`   |
+
+### Exemptions (mode: `exempt-only`, default)
+
+Some links are intentionally one-way. Defaults:
+
+- **`outputs/**`** — ephemeral artifacts
+- **External URLs** (`*://*`) — out of wiki scope
+
+Anything outside an exemption glob must be bidirectional.
+
+---
+
+## Log Format
+
+Append-only. One line per skill invocation. Format:
+
+```markdown
+## [YYYY-MM-DD] skill | details
+```
+
+`grep "^## \[" wiki/log.md | tail -10` gives you recent activity.
+
+---
+
+## Graph
+
+`wiki/graph/edges.jsonl` and `wiki/graph/citations.jsonl` are auto-generated. Never edit manually. The full set of edge types lives in `_lumina/scripts/schemas.mjs` — open it when you need to pick a type or check what is allowed.
+
+---
+
+## Constraints (Non-Negotiable)
+
+- **`raw/` is user-owned**: never modify or delete existing files; additions only via the two named paths above.
+- **`graph/` is auto-generated**: only modify via the graph rebuild step.
+- **Bidirectional links are mandatory**: forward link and reverse link in the same operation.
+- **`index.md` updated on every ingest**: every new page must be cataloged immediately.
+- **`log.md` is append-only**: never rewrite history.
+- **Skill flags are user-owned**: never invent, flip, or drop a flag based on repo state alone. If the user omitted a parameter, only fill it in when the skill explicitly documents a default; otherwise ask.
+- **No silent overwrites**: preserve sections marked with `<!-- user-edited -->` comment.
+- **Cite when uncertain**: link the source explicitly for low-confidence claims.
+
+---
+
+## Skills
+
+Skills live in `.agents/skills/` and are invoked via slash commands. Active install recorded in `_lumina/manifest.json`.
+
+### Core skills (always present)
+
+| Skill         | Trigger        | What it does                                          |
+|---------------|---------------|-------------------------------------------------------|
+| `/lumi-init`   | manual, first  | Bootstrap wiki from existing `raw/` content          |
+| `/lumi-ingest` | manual         | Read a source and write a wiki page. It asks you to review the draft, then continues on its own unless something needs your judgment |
+| `/lumi-ask`    | manual         | Query wiki, synthesize answer, optionally file page   |
+| `/lumi-edit`   | manual         | Add/remove/revise wiki content per user request       |
+| `/lumi-check`  | manual/weekly  | Lint: broken links, orphans, missing reverse links    |
+| `/lumi-reset`  | manual         | Scoped destructive cleanup                            |
+| `/lumi-verify` | manual         | Check that wiki pages match the sources they cite; reports suspicious statements for the user to review; never auto-edits                            |
+
+
+---
+
+## Tooling Conventions
+
+- **`_lumina/scripts/lint.mjs`** — pure-Node markdown linter, runs offline.
+- **`_lumina/scripts/wiki.mjs`** — wiki engine (frontmatter, graph mutation, slug, log).
+- **`_lumina/scripts/reset.mjs`** — scoped destructive reset.
+- **`_lumina/tools/extract_pdf.py`** — PDF text extractor (pypdf-based); used by `/lumi-ingest` and `/lumi-reading-chapter-ingest` when the host IDE cannot read PDFs natively.
+- **`_lumina/tools/fetch_pdf.py`** — URL → `raw/download/<resource>/` PDF downloader (streaming, atomic, idempotent); used by `/lumi-ingest` Mode B when the input is a URL or paper identifier.
+- **`_lumina/tools/requirements.txt`** — Python dependencies for bundled tools. Run `pip install -r _lumina/tools/requirements.txt` when a tool reports a missing package.
+
+---
+
+## How To Use This Wiki (For New LLM Sessions)
+
+1. Read this file (you are doing it now).
+2. Read `wiki/index.md` to learn what already exists.
+3. Read `wiki/log.md`'s last 20 entries to learn what happened recently.
+4. When the user invokes a skill, read the skill's `SKILL.md` first.
+5. When in doubt about page structure, open `_lumina/schema/page-templates.md`.
+6. When in doubt about scope, ask the user — never silently expand it.
+
+The wiki is a long-running collaboration. Maintain it patiently.
+
+<!-- /lumina:schema -->
+
+---
+
+## Contributors
+
+Thanks to everyone who has contributed to Lumina Wiki!
+
+[![Contributors](https://contrib.rocks/image?repo=tronghieu/lumina-wiki)](https://github.com/tronghieu/lumina-wiki/graphs/contributors)
+
+**Want to contribute?** Read [CONTRIBUTING.md](CONTRIBUTING.md) to get started — bug reports, new skills, tool integrations, and translations are all welcome.

package/README.vi.md

@@ -184,6 +184,7 @@ Xem [Hướng dẫn Nâng cao](docs/user-guide/advanced-qmd.vi.md) để biết
 | | `/lumi-research-prefill` | Tạo trước các khái niệm nền tảng để tránh trùng lặp. |
 | | `/lumi-research-topic` | Gom các khái niệm và nguồn liên quan trong wiki thành một trang chủ đề tại `wiki/topics/<slug>.md`. AI đề xuất danh sách để bạn xem và xác nhận trước khi trang được tạo. Dùng sau khi đã nạp nhiều tài liệu và muốn tổng hợp một nhóm ý tưởng thành trang riêng. |
 | | `/lumi-research-setup` | Giúp cấu hình API key cho các công cụ nghiên cứu. |
+| | `/lumi-research-watch-run` | Chạy một lượt khám phá theo lịch dựa trên watchlist (chủ đề + nguồn RSS / Atom). Chỉ chạy khi bạn yêu cầu. |
 | **Reading** | `/lumi-reading-chapter-ingest`| Nạp kiến thức sách theo từng chương. |
 | | `/lumi-reading-character-track`| Theo dõi các nhân vật và mối quan hệ của họ trong truyện. |
 | | `/lumi-reading-theme-map` | Xác định và lập bản đồ các chủ đề trong một câu chuyện. |
@@ -204,8 +205,8 @@ Lumina-Wiki đang phát triển nhanh chóng. Dưới đây là lộ trình hư
 - [x] **Cài đặt đa ngôn ngữ:** Chọn Tiếng Anh, Tiếng Việt hoặc Tiếng Trung làm ngôn ngữ chính khi cài đặt. *(đã phát hành trong v1.2)*
 - [x] **Nạp DOCX, RTF & EPUB native:** Đưa thẳng file Word, Rich Text và sách EPUB vào wiki qua research pack. *(đã phát hành trong v1.x)*
 - [x] **Cải thiện CI/CD:** Hỗ trợ chính thức cho môi trường Bun và Node 22. *(đã phát hành trong v1.2)*
-- [ ] **Mở rộng nguồn dữ liệu toàn cầu:** Tích hợp trực tiếp với OpenAlex, CORE và Unpaywall để tra cứu DOI-to-PDF đáng tin cậy.
-- [ ] **Theo dõi RSS & Blog:** Tự động phát hiện bài báo mới từ các blog phòng thí nghiệm và tạp chí yêu thích.
+- [x] **Mở rộng nguồn dữ liệu toàn cầu:** Tích hợp trực tiếp với OpenAlex, CORE và Unpaywall để tra cứu DOI-to-PDF đáng tin cậy. *(ra mắt trong v1.6)*
+- [x] **Theo dõi RSS & Blog:** Tự động phát hiện bài báo mới từ các blog phòng thí nghiệm và tạp chí yêu thích qua các mục `type: feed` trong watchlist. *(ra mắt trong v1.6)*
 - [ ] **Xếp hạng bài báo nâng cao:** Xem điểm số ảnh hưởng và tín hiệu chất lượng cho các nghiên cứu của bạn.
 
 **Dài hạn (Nghiên cứu sâu & Tích hợp)**
@@ -249,3 +250,13 @@ npm run test:all
 - [简体中文 (Tiếng Trung)](README.zh.md)
 
 **Giấy phép:** [MIT](LICENSE) © Lưu Trọng Hiếu.
+
+---
+
+## Người đóng góp
+
+Cảm ơn tất cả những người đã đóng góp cho Lumina Wiki!
+
+[![Contributors](https://contrib.rocks/image?repo=tronghieu/lumina-wiki)](https://github.com/tronghieu/lumina-wiki/graphs/contributors)
+
+**Muốn đóng góp?** Đọc [CONTRIBUTING.md](CONTRIBUTING.md) để bắt đầu — báo lỗi, thêm skill mới, tích hợp công cụ, hay dịch thuật đều được chào đón.

package/README.zh.md

@@ -185,6 +185,7 @@ npx skills add https://github.com/tobi/qmd --skill qmd
 | | `/lumi-research-prefill` | 预先生成基础概念，避免重复。 |
 | | `/lumi-research-topic` | 把 wiki 中已有的相关概念和来源汇聚成一个主题页，保存在 `wiki/topics/<slug>.md`。AI 会提议收录哪些内容，由你确认后再生成页面。多次 `/lumi-ingest` 之后，用它把一组相关想法整理成独立的主题页。 |
 | | `/lumi-research-setup` | 帮助配置研究工具的 API key。 |
+| | `/lumi-research-watch-run` | 基于 watchlist 运行一次计划式发现（主题 + RSS / Atom 源）。仅在你要求时才运行。 |
 | **Reading** | `/lumi-reading-chapter-ingest`| 按章节导入书籍知识。 |
 | | `/lumi-reading-character-track`| 追踪故事中的角色及其关系。 |
 | | `/lumi-reading-theme-map` | 识别并映射故事主题。 |
@@ -205,8 +206,8 @@ Lumina-Wiki 正在快速演进。这是我们的用户路线图：
 - [x] **多语言安装：** 安装时可选英文、越南文或中文作为主语言。*（v1.2 已发布）*
 - [x] **原生 DOCX、RTF 与 EPUB 导入：** 通过 research pack 将 Word、Rich Text 与 EPUB 电子书直接导入维基。*（v1.x 已发布）*
 - [x] **改进的 CI/CD：** 正式支持 Bun 和 Node 22 环境。*（v1.2 已发布）*
-- [ ] **全球数据源扩展：** 直接集成 OpenAlex、CORE 和 Unpaywall，实现可靠的 DOI-to-PDF 解析。
-- [ ] **RSS 与博客监控：** 自动从您喜爱的实验室博客和期刊中发现新论文。
+- [x] **全球数据源扩展：** 直接集成 OpenAlex、CORE 和 Unpaywall，实现可靠的 DOI-to-PDF 解析。*（在 v1.6 中发布）*
+- [x] **RSS 与博客监控：** 通过 watchlist 中的 `type: feed` 项，自动从您喜爱的实验室博客和期刊中发现新论文。*（在 v1.6 中发布）*
 - [ ] **高级论文排名：** 查看研究论文的影响力评分和质量信号。
 
 **长期计划（深度研究与集成）**
@@ -251,3 +252,13 @@ npm run test:all
 - [Tiếng Việt（越南语）](README.vi.md)
 
 **许可：** [MIT](LICENSE) © Lưu Trọng Hiếu.
+
+---
+
+## 贡献者
+
+感谢所有为 Lumina Wiki 做出贡献的人！
+
+[![Contributors](https://contrib.rocks/image?repo=tronghieu/lumina-wiki)](https://github.com/tronghieu/lumina-wiki/graphs/contributors)
+
+**想要贡献？** 请阅读 [CONTRIBUTING.md](CONTRIBUTING.md) 开始参与——欢迎提交 Bug 报告、新 skill、工具集成以及翻译。

package/package.json

@@ -1,7 +1,7 @@
 {
   "$schema": "https://json.schemastore.org/package.json",
   "name": "lumina-wiki",
-  "version": "1.5.0",
+  "version": "1.6.0",
   "description": "Domain-agnostic, multi-IDE wiki scaffolder — Karpathy's LLM-Wiki vision, cross-platform and pack-based.",
   "keywords": [
     "llm-wiki",

package/src/installer/commands.js

@@ -952,26 +952,25 @@ function buildIdeStub(target, vars) {
 async function copyScripts(projectRoot) {
   const destDir = join(projectRoot, '_lumina', 'scripts');
   const scriptFiles = ['wiki.mjs', 'lint.mjs', 'reset.mjs', 'schemas.mjs', 'discover-runner.mjs', 'external-ids.mjs', 'parse-ids.mjs', 'merge-ids.mjs', 'build-source.mjs'];
-  for (const file of scriptFiles) {
-    const src = join(SCRIPTS_DIR, file);
-    const dest = join(destDir, file);
+  // Parallel: each copy is independent; destDir is created earlier.
+  await Promise.all(scriptFiles.map(async file => {
     try {
-      await copyFile(src, dest);
+      await copyFile(join(SCRIPTS_DIR, file), join(destDir, file));
     } catch (_) {
       // Scripts may not exist yet (P4+ work); skip gracefully
     }
-  }
+  }));
+  // Ensure the lib subdir once, then parallelize the file copies.
+  const libDir = join(destDir, 'lib');
+  await ensureDir(libDir);
   const libFiles = ['watchlist-config.mjs', 'discovery-state.mjs'];
-  for (const file of libFiles) {
-    const src = join(SCRIPTS_DIR, 'lib', file);
-    const dest = join(destDir, 'lib', file);
+  await Promise.all(libFiles.map(async file => {
     try {
-      await ensureDir(dirname(dest));
-      await copyFile(src, dest);
+      await copyFile(join(SCRIPTS_DIR, 'lib', file), join(libDir, file));
     } catch (_) {
       // Script libs may not exist yet; skip gracefully
     }
-  }
+  }));
 }
 
 async function copyChangelog(projectRoot) {
@@ -1052,6 +1051,7 @@ function getSkillDefs(packs) {
       { name: 'setup',    canonicalId: 'lumi-research-setup',    displayName: '/lumi-research-setup' },
       { name: 'topic',     canonicalId: 'lumi-research-topic',     displayName: '/lumi-research-topic' },
       { name: 'watchlist', canonicalId: 'lumi-research-watchlist', displayName: '/lumi-research-watchlist' },
+      { name: 'watch-run', canonicalId: 'lumi-research-watch-run', displayName: '/lumi-research-watch-run' },
     ];
     for (const s of researchSkills) {
       defs.push({ ...s, pack: 'research', srcPackPath: 'packs/research' });
@@ -1086,19 +1086,22 @@ async function copyTools(projectRoot, { research }) {
   const destDir = join(projectRoot, '_lumina', 'tools');
   const coreTools = ['extract_pdf.py', 'fetch_pdf.py', 'id_utils.py'];
   const researchTools = [
-    '_env.py', 'discover.py', 'init_discovery.py', 'prepare_source.py',
+    '_env.py', '_cache.py', 'discover.py', 'init_discovery.py', 'prepare_source.py',
     'fetch_arxiv.py', 'fetch_wikipedia.py', 'fetch_s2.py', 'fetch_deepxiv.py',
+    'fetch_openalex.py', 'fetch_unpaywall.py', 'fetch_core.py', 'resolve_pdf.py',
+    'fetch_rss.py',
   ];
   const toolFiles = research ? [...coreTools, ...researchTools] : coreTools;
-  for (const file of toolFiles) {
-    const src = join(TOOLS_DIR, file);
-    const dest = join(destDir, file);
+  // Parallelize: each copy is independent and destDir already exists.
+  // Sequential awaits were the main Windows cold-start regression in v1.4
+  // (~30 ms per file × 14 files dominates on NTFS + Defender).
+  await Promise.all(toolFiles.map(async file => {
     try {
-      await copyFile(src, dest);
+      await copyFile(join(TOOLS_DIR, file), join(destDir, file));
     } catch (_) {
       // Tool not yet authored; skip
     }
-  }
+  }));
   try {
     await copyFile(join(TOOLS_DIR, 'requirements.txt'), join(destDir, 'requirements.txt'));
   } catch (_) {
@@ -1110,7 +1113,9 @@ async function renderSchemaDocs(projectRoot, templateVars) {
   const schemaDir = join(projectRoot, '_lumina', 'schema');
   const schemaDocs = ['page-templates.md', 'cross-reference-packs.md', 'graph-packs.md', 'lumi-help.csv', 'lumi-help-runbook.md'];
 
-  for (const doc of schemaDocs) {
+  // Parallel render + atomicWrite. Each doc is independent; schemaDir is
+  // already created by the dirsToCreate loop in installCommand.
+  await Promise.all(schemaDocs.map(async doc => {
     const templatePath = join(TEMPLATES_DIR, '_lumina', 'schema', doc);
     const destPath = join(schemaDir, doc);
     let content;
@@ -1121,7 +1126,7 @@ async function renderSchemaDocs(projectRoot, templateVars) {
       content = `# ${doc}\n\n_This file is managed by the Lumina installer._\n`;
     }
     await atomicWrite(destPath, content);
-  }
+  }));
 }
 
 async function renderEnvExample(projectRoot) {

package/src/installer/prompts.js

@@ -151,7 +151,7 @@ export function buildPromptList(existingManifest, defaultLocale = 'en') {
  * @param {Function} [opts.t]                   - Locale translator function.
  * @returns {Promise<InstallAnswers>}
  */
-export async function runInstallPrompts({ acceptDefaults = false, cwd = process.cwd(), existingManifest = null, defaultLocale = 'en', t = null } = {}) {
+export async function runInstallPrompts({ acceptDefaults = false, cwd = process.cwd(), existingManifest = null, defaultLocale = 'en', t: initialT = null } = {}) {
   if (acceptDefaults) {
     const loc = existingManifest?.locale ?? defaultLocale;
     return defaultAnswers(cwd, loc);
@@ -159,6 +159,11 @@ export async function runInstallPrompts({ acceptDefaults = false, cwd = process.
 
   const { intro, outro, text, multiselect, select, confirm, isCancel, cancel } = await getClack();
 
+  // t starts as whatever the caller passed (may be null on fresh install where
+  // locale is unknown). After Prompt 0 selects the locale, we rebind t to the
+  // user-chosen locale so prompts 1-5 are localized.
+  let t = initialT;
+
   // t may not be available yet at intro time (locale not yet selected);
   // use hardcoded EN for intro — this is the chicken-and-egg prompt.
   intro('Lumina Wiki Installer');
@@ -179,14 +184,26 @@ export async function runInstallPrompts({ acceptDefaults = false, cwd = process.
   const locale = localeRaw;
   const langDefault = LOCALE_LANGUAGE_NAME[locale] ?? 'English';
 
+  // Rebind t to the just-selected locale so the remaining prompts render in
+  // the user's chosen language. Without this, prompts 1-5 silently fall back
+  // to EN literals even when the user picked vi/zh at Prompt 0.
+  try {
+    const { loadLocale } = await import('./locales.js');
+    const localeMod = await loadLocale(locale);
+    t = localeMod.t;
+  } catch {
+    // Keep whatever t we had (possibly null → EN literals); never block install.
+  }
+
   // Interactive locale-switch confirmation (Phase 5 §60-63):
   // If user picks a locale different from the installed one on an upgrade,
   // require explicit confirmation before destructively rewriting README.md
   // and IDE stubs. Default N — protects user content.
   if (existingManifest?.locale && existingManifest.locale !== locale) {
     const proceed = await confirm({
-      // Use trilingual literal — t() not yet bound to chosen locale and the
-      // user is mid-switch; both old and new locales are relevant context.
+      // Use trilingual literal — user is mid-switch, so both old and new
+      // locales are relevant context. t() is available here but intentionally
+      // not used so the warning is legible in either locale.
       message: `Locale change ${existingManifest.locale} -> ${locale} will rewrite README.md and IDE stubs in the new locale. Outside-schema edits are preserved. Continue?`,
       initialValue: false,
     });