dsh-skill-recommender
DeepSeek Harness 的技能推荐器：浏览本地 Codex/Claude/DSH 会话记录，构建用户画像，并通过可调权重的匹配指数推荐合适的开源技能（指数越高仅表示相关性越高）。提供工具和 Web 设置面板。
插件会安装到这里；不确定时保持 web。
npx -y @deepseek-ai/dsh plugin --profile web add dsh-skill-recommender@0.1.5Skill Recommender 以 dsh-skill-recommender 发布，当前版本为 0.1.5。Plugin Hub 会校验它的 manifest，并保存精确安装来源，便于复现安装结果。
A skill recommender for DeepSeek Harness. It browses your local session records across Codex, Claude and DSH, builds a compact user profile, ingests an open-source skill catalog, and ranks skills with a weighted score gated by a tunable match index — the higher the index, the higher-relevance the recommendations only.
~/.dsh/sessions/**/session.jsonl.zstd), Codex (~/.codex/*.jsonl), Claude (~/.claude/projects/**/*.jsonl) — no API keys, purely local file reads.~/.agents/skills, ~/.dsh/skills, your Obsidian 2️⃣ AI/Skill) + remote awesome lists (awesome-dsh-skills, awesome-dsh-plugin, awesome-deepseek-harness, Claude skills ecosystem), plus a small bundled seed so it always returns something.score = Σ(w_i × sim_i) / Σ(w_i), where dimensions are topic / tool / task / proximity-to-installed. A global match index (0–100, default 60) is the gate: only skills with score ≥ index are returned, then top-N.dsh-skill-studiodsh-skill-studio extracts reusable skills from your own sessions. This plugin recommends third-party open-source skills by extrapolating your profile. They complement each other.
| Tool | Purpose |
|---|---|
recommender_scan | Scan sessions, build the profile, produce an initial recommendation run. |
recommender_recommend | Recommend open-source skills; optionally override index / topN / weights on the fly. |
recommender_profile | Show the current user profile (topics, tools, tasks, projects). |
recommender_config | Configure sources, catalogs, window, match index, per-dimension weights, optional LLM enrichment. |
recommender_status | Show plugin status without leaking secrets. |
A Skill 推荐器 card in the Web settings page: scan button, live 匹配指数 slider (0–100), four per-dimension weight sliders (topic / tool / task / proximity), catalog toggles, and recommendation cards with scores, per-dimension breakdowns and a GitHub link. The index + weights are saved to ~/.dsh/dsh-skill-recommender/config.json (mode 0600).
Requires DeepSeek Harness ≥ 0.1.5-rc.1 (declared as dsh.engines.dsh in the package manifest, so the DSH plugin marketplace can report it) and is verified against 0.1.5-rc.1. This build carries the DSH 0.1.5 adaptations: the strict tool-result contract (lossless-JSON snapshot, additionalProperties: false schema validation, and output.render returning ContentBlock[]) plus executable resolution that survives a launchd-started host whose PATH is only /usr/bin:/bin. Since this version the peerDependencies range explicitly declares compatibility with DSH 0.2.0-rc.2 (^0.2.0-rc.2 is now included), so no compatibility warning appears on 0.2.0-rc.2. No functional change.
Add the plugin to your profile:
# from npm
dsh plugin --profile web add dsh-skill-recommender
# or from GitHub (the repo carries the `dsh-plugin` topic)
dsh plugin --profile web add github:zhengjy01/dsh-skill-recommender
# local development
dsh plugin --profile web add link:/path/to/dsh-skill-recommender
Expected: the command prints the resolved package and records it in that
profile's dsh.profile.bundles.
Screenshot slot 1 — install output. Capture the terminal right after the command, last ~10 lines (package + profile). Redact your username / home path. Save as
docs/images/dsh-skill-recommender-1-install.png, then replace this block with.
Restart the host (tools + routes), then hard-refresh the browser page (client
panel). Config key: skill-recommender in the bundle patch layer.
Expected: 设置 (Settings) shows the Skill 推荐器 card.
Screenshot slot 2 — the settings card. Capture the card's first screen (scan button + index slider). Redact any catalog names you consider private. Save as
docs/images/dsh-skill-recommender-2-card.png.
Press 扫描 / Scan.
Expected: the progress bar advances, then recommendation cards appear with a score, the per-dimension breakdown, and a clickable source link.
Screenshot slot 3 — recommendations with scores. Capture the result list (2–3 cards are enough). Save as
docs/images/dsh-skill-recommender-3-results.png.
Optional: drag the 匹配指数 / index slider up and re-scan.
Expected: fewer, higher-relevance recommendations (index 90 typically
returns very few).
Screenshot slot 4 — the index gate works. Capture the list before/after raising the index. Save as
docs/images/dsh-skill-recommender-4-index.png.
pnpm install
pnpm bundle # builds lib/index.js (ESM) + lib/client.js (browser bundle)
node tests/smoke.mjs
{topic: 50, tool: 50, task: 35, near: 40}; the index gate defaults to 60.github-discovery source queries the GitHub search API across the entire platform (agent/claude/codex skills, SKILL.md, awesome skills — 星标排序), so recommendations are not limited to one designated list. Results are cached (6h TTL); the unauthenticated search limit (10 req/min) is respected.skill or plugin. By default only skills are recommended (types: ['skill']); tick 含插件 in the panel to also include DSH plugins. Both markdown lists and tables are parsed (DSH skill catalogs use tables).autoScanMinutes (default 60, 0 = off) in the panel — the host refreshes the cached result in the background, and the panel shows the last result instantly on open. The cache is persisted to ~/.dsh/dsh-skill-recommender/last-result.json (0600).MIT
| # | Where / 放在哪 | What it shows / 展示什么 | How to capture / 怎么截 | Suggested filename / 建议文件名 |
|---|---|---|---|---|
| 1 | Install | The install command's output (package + profile) | Terminal right after step 1, last ~10 lines; redact home path | docs/images/dsh-skill-recommender-1-install.png |
| 2 | Install | The Skill 推荐器 card in Settings | Settings page, card header + scan button | docs/images/dsh-skill-recommender-2-card.png |
| 3 | Usage | Recommendation cards with scores | Result list, 2–3 cards | docs/images/dsh-skill-recommender-3-results.png |
| 4 | Usage | Raising the index narrows the list | Same list before/after raising the index | docs/images/dsh-skill-recommender-4-index.png |
After adding the images, re-run npm pack --dry-run and the portability gate
(npm run verify) — the README is part of the published tarball.
补图后请重跑 npm pack --dry-run 与可移植性门禁（npm run verify）。