dsh-skill-hub 0.3.27 → 0.3.28

package/CONTRIBUTING.md

@@ -39,7 +39,7 @@ src/client/locales/     dictionaries by view (common/skills/market/sources/detai
 ```bash
 npm install
 npm run typecheck   # tsc --noEmit
-npm test            # vitest (295 tests across 16 suites)
+npm test            # vitest (366 tests across 20 suites)
 npm run build       # tsc declarations + tsdown bundles (lib/index.js + lib/client.js)
 ```
 

package/README.md

@@ -36,9 +36,13 @@ Two surfaces — the hub panel lives at **Settings → 技能**, and this plugin
 - **Toggle** — per-skill switches and per-group tri-state switches with a conflict dialog (close all / keep on). Disabling renames the discovery file (never deletes); disabled skills stay inspectable and re-enableable from their detail page. Only `~/.dsh/skills` & `~/.agents/skills` are writable; everything else is read-only.
 - **Organize** — scenes (tags) plus auto-aggregated source collections, all drag-reorderable and persisted in `~/.dsh/dsh-skill-hub.json`. Edit mode reveals delete/reorder without cluttering the read view.
 - **Diagnose & fix** — files the provider skips (missing frontmatter, bad YAML, name mismatch, short description) show up with reasons; auto-fixable ones (e.g. unquoted `:` in descriptions) get a one-click Fix button.
+- **Content audit** — enabled skills' bodies (`SKILL.md`) are checked too: raw ESC control characters, Unicode tag characters, and bidirectional controls (these three make "reviewed" not equal "seen" — a market import carrying them is refused outright), zero-width characters, `curl … | sh` remote execution, SSH/cloud-credential/keychain paths, over-long descriptions, and the 2 MiB per-file / 10 MiB per-skill / 500-file limits. Pure local static matching: no model call, no network. A market import scans the whole bundle (including `scripts/`); the standing check only reads `SKILL.md` — it runs on every 5-second poll, and walking every directory would slow the panel down.
+- **Announcement budget** — the name + description total that the first progressive-disclosure layer injects comes back with the catalog, so the panel can show characters and an estimated token count; past 8000 characters it turns red (clients truncate the tail there, which shows up as "installed but never triggers").
 - **Scaffold** — new-skill wizard writing to `~/.dsh/skills` or `~/.agents/skills` (`SKILL.md` template below).
-- **Market** — built-in curated repos plus custom `owner/repo` sources. Any top-level directory containing `SKILL.md` scans as a root (no allowlist), and a `SKILL.md` at the repo root itself scans as a single skill named after the repo (minus repo tooling such as `.github/`). Async import with byte-level progress and cancel. Each source pins a version — click the ref badge to switch between releases, branches, or a custom ref.
+- **Market** — built-in curated repos plus custom `owner/repo` sources. Any top-level directory containing `SKILL.md` scans as a root (no allowlist), and a `SKILL.md` at the repo root itself scans as a single skill named after the repo (minus repo tooling such as `.github/`). Async import with byte-level progress and cancel. Each source pins a version — click the ref badge to switch between releases, branches, or a custom ref. Every import passes the content audit first: a skill with a blocking finding is refused as a whole and its temp directory is removed (nothing left on disk).
 - **Track updates** — imported skills record a repo + commit snapshot. Check all / update-all, per-source badges (installed / updatable / deleted upstream / new release). Sync overwrites local edits (with confirm); upstream deletions move into a restorable trash that keeps source and scene membership.
+- **Local-edit detection** — a content hash of the whole skill (path + size + body) is sampled right after the bytes land and compared on every check: skills that no longer match get an "edited locally" badge, and the sync dialog marks exactly which ones the overwrite will discard. Records without a hash are treated as unknown, never as edited.
+- **Import from another agent** — read-only scan of the user-level skill folders of Claude Code / Codex / Cursor / Gemini CLI / Copilot / OpenCode / Goose, then copy the selected ones into `~/.dsh/skills`. The originals are untouched; flat skills land as directory bundles; the same content audit and size caps apply.
 - **Stats** — per-skill call counts + last-used times from session logs (incremental cache), group summaries; window and scan interval live-configurable from the settings card.
 - **Settings card** — registered on the plugin's own page in the **Plugins manager** (sidebar → 插件 → dsh-skill-hub; collapsed until you expand the title): master switch, announce-to-agent, invocation dot colors, usage display toggles, stats window/interval.
 
@@ -51,6 +55,8 @@ Two surfaces — the hub panel lives at **Settings → 技能**, and this plugin
 | Catalog | user roots, self-scanned | `ctx.skills` registry, all roots + third-party |
 | Toggle | ❌ | ✅ per-skill + per-group, never deletes |
 | Diagnostics | ❌ | ✅ reasons + one-click fix |
+| Content audit | ❌ | ✅ hidden characters / dangerous commands / credential paths, blocked at import |
+| Cross-agent import | ❌ | ✅ read-only scan of 7 tool folders, pick and adopt |
 | Market | ❌ | ✅ built-in + custom, version pins, update-all |
 | Source tracking | ❌ | ✅ check/sync/trash with restore |
 | Stats | ❌ | ✅ counts + last-used |
@@ -101,6 +107,8 @@ Host uses only `ctx.skills.snapshot/get`, `ctx.webServer.register`, `ctx.systemP
 | `/api/skill-hub/repo/import/progress?jobId=` | GET | poll job |
 | `/api/skill-hub/repo/import/cancel` | POST | cancel job |
 | `/api/skill-hub/sources` etc. | GET/POST | list/check/sync/delete/restore/clear trash |
+| `/api/skill-hub/foreign` | GET | read-only scan of other agents' user-level skill folders |
+| `/api/skill-hub/foreign/import` | POST | `{items:[{agent,name}]}` copied into `~/.dsh/skills` |
 | `/api/skill-hub/update` | GET | plugin latest release |
 | unknown `/api/skill-hub/*` | GET/POST | `404 {error}` naming the path — a typo must not read as an auth failure (the SPA fallback answers 401 for `/api/*` it does not own) |
 
@@ -108,7 +116,7 @@ Host uses only `ctx.skills.snapshot/get`, `ctx.webServer.register`, `ctx.systemP
 
 ```bash
 npm run typecheck  # tsc --noEmit
-npm test           # 295 tests, 16 suites
+npm test           # 366 tests, 20 suites
 npm run build      # tsc + tsdown → lib/index.js + lib/client.js
 ```
 

package/README.zh.md

@@ -36,9 +36,13 @@ dsh plugin --profile web add dsh-skill-hub
 - **开关** — 单技能开关 + 整组三态开关，跨组冲突弹窗（全部关闭/保留开启）。禁用只重命名发现文件（不删除）；禁用的技能仍可查看正文、可一键重新启用。仅 `~/.dsh/skills` 与 `~/.agents/skills` 可写，其余只读。
 - **整理** — 场景（tag）+ 自动聚合的来源集合，全部可拖拽排序并持久化到 `~/.dsh/dsh-skill-hub.json`。编辑模式收敛删除/排序控件，阅读视图保持干净。
 - **诊断与修复** — provider 跳过的文件给出原因（缺 frontmatter、YAML 非法、名称不一致、描述过短）；可自动修复的（如描述里未加引号的 `:`）一键 Fix 落盘。
+- **内容审计** — 已启用技能的正文（`SKILL.md`）也体检：裸 ESC 控制字符、Unicode 标签字符、双向控制符（这三类会让「看过」不等于「看见」，市场导入时直接拒绝落盘）、零宽字符、`curl … | sh` 式远程执行、SSH/云凭据/钥匙串路径、超长 description，以及单文件 2 MiB / 单技能 10 MiB / 500 文件上限。规则是纯本地静态匹配，不跑模型、不出网。市场导入会扫整包（含 `scripts/`），常驻体检只扫 `SKILL.md`（每 5 秒轮询，扫全目录会拖慢面板）。
+- **公告预算** — 一层渐进加载注入的 name + description 总量随目录一起返回，面板显示字符数与 token 估算；超过 8000 字符变红提示（超出后客户端会截断尾部条目，表现为「装了却不触发」）。
 - **新建** — 新技能向导，写入 `~/.dsh/skills` 或 `~/.agents/skills`（`SKILL.md` 模板见下）。
-- **市场** — 内置精选仓库 + 自定义 `owner/repo`。任何含 `SKILL.md` 的顶层目录都可扫描（无白名单）；`SKILL.md` 直接放在仓库根时，整个仓库算一个技能，技能名取仓库名（`.github/` 等仓库基建不计入）。异步导入，字节级进度+取消。每个源钉一个版本 —— 点 ref 徽标可在发布版/分支/手输之间切换。
+- **市场** — 内置精选仓库 + 自定义 `owner/repo`。任何含 `SKILL.md` 的顶层目录都可扫描（无白名单）；`SKILL.md` 直接放在仓库根时，整个仓库算一个技能，技能名取仓库名（`.github/` 等仓库基建不计入）。异步导入，字节级进度+取消。每个源钉一个版本 —— 点 ref 徽标可在发布版/分支/手输之间切换。导入前先过内容审计，命中阻断级规则的技能整包拒绝，临时目录一并清掉（磁盘不留痕）。
 - **跟踪更新** — 导入的技能记录 repo+commit 快照。检查全部/一键全更；每源徽章（已装/可更新/上游已删/新版本）。同步覆盖本地修改（先确认）；上游删除跟进移入回收站，恢复保留来源与场景归属。
+- **本地改动检测** — 落盘后立刻采集整个技能的内容哈希（路径+大小+内容），检查时与记录比对：不一致的技能挂「本地改动 N 个」徽章，同步确认弹窗里逐个标出会被覆盖的那几个。旧记录没有哈希时按「未知」处理，不误报。
+- **从其他 Agent 导入** — 只读扫描 Claude Code / Codex / Cursor / Gemini CLI / Copilot / OpenCode / Goose 的用户级技能目录，勾选后复制到 `~/.dsh/skills`。原目录不受影响；扁平技能统一落成目录包；同样过内容审计与体积上限。
 - **统计** — 会话日志的调用次数+最近使用，分组头汇总；窗口与扫描间隔在设置卡片实时可调。
 - **设置卡片** — 注册在 **插件管理页** 的插件自有页面上（侧边栏「插件」→ dsh-skill-hub，默认折叠，点标题展开）：总开关、向 Agent 公告、调用圆点颜色、用量显示开关、统计窗口/间隔。
 
@@ -51,6 +55,8 @@ dsh plugin --profile web add dsh-skill-hub
 | 目录 | 自扫盘、仅用户根 | `ctx.skills` 全量+第三方 |
 | 开关 | ❌ | ✅ 单技能+整组，从不删除 |
 | 诊断 | ❌ | ✅ 原因+一键修复 |
+| 内容审计 | ❌ | ✅ 隐藏字符/危险命令/凭据路径，导入即拦 |
+| 跨 agent 导入 | ❌ | ✅ 只读扫描 7 家目录，勾选收编 |
 | 市场 | ❌ | ✅ 内置+自定义，版本钉选，一键全更 |
 | 来源跟踪 | ❌ | ✅ 检查/同步/回收站可恢复 |
 | 统计 | ❌ | ✅ 次数+最近使用 |
@@ -101,6 +107,8 @@ GitHub 仓库 ──扫描/导入──▶ ~/.dsh/skills
 | `/api/skill-hub/repo/import/progress?jobId=` | GET | 轮询进度 |
 | `/api/skill-hub/repo/import/cancel` | POST | 取消任务 |
 | `/api/skill-hub/sources` 等 | GET/POST | 来源 列表/检查/同步/删除/恢复/清空回收站 |
+| `/api/skill-hub/foreign` | GET | 其他 agent 用户级技能目录的只读扫描 |
+| `/api/skill-hub/foreign/import` | POST | `{items:[{agent,name}]}` 收编到 `~/.dsh/skills` |
 | `/api/skill-hub/update` | GET | 插件最新发布 |
 | 未知的 `/api/skill-hub/*` | GET/POST | `404 {error}` 且写明路径 —— 拼错路径不该被读成鉴权失败（宿主 SPA fallback 对自己不拥有的 `/api/*` 一律回 401） |
 
@@ -108,7 +116,7 @@ GitHub 仓库 ──扫描/导入──▶ ~/.dsh/skills
 
 ```bash
 npm run typecheck  # tsc --noEmit
-npm test           # 295 tests, 16 suites
+npm test           # 366 tests, 20 suites
 npm run build      # tsc + tsdown → lib/index.js + lib/client.js
 ```