{"skill":{"slug":"skill-picker","displayName":"Skill Picker","summary":"Discovers and recommends the best agent skills based on user intent — not just keywords. Activates when users ask \"how do I do X\", \"find a skill for X\", \"is...","description":"---\nname: skill-picker\ndescription: Discovers and recommends the best agent skills based on user intent — not just keywords. Activates when users ask \"how do I do X\", \"find a skill for X\", \"is there a skill that...\", or whenever a task would benefit from a specialized skill. Also activates proactively when the agent detects repeated manual tasks, workflow gaps, or missing capabilities mid-conversation. Supports English and Chinese. Recommends skill combinations, not just single skills. This skill only suggests install commands — it never executes installs autonomously. All installs require explicit user confirmation.\ntools:\n  - npx\nbinaries:\n  - npx\n---\n\n# Skill Picker\n\nDiscover and recommend the right skills at the right moment.\n\n> **What this skill does with `npx`**: It runs `npx skills find [query]` to search the skills registry and display results. It does NOT run `npx skills add` on its own — install commands are always shown to the user for manual execution or explicit confirmation.\n\n---\n\n## Core Philosophy\n\n**Don't match keywords. Understand intent.**\n\nThe difference:\n- Keyword match: User says \"GitHub\" → search \"github\"\n- Intent match: User says \"我的PR没人review\" → understand they need async code review automation → recommend `gh-issues` + `coding-agent` as a combination\n\nAlways ask: *What outcome does the user actually want?* Then find the shortest path to that outcome using available skills.\n\n> **Security principle**: `npx skills find` is used for searching only. `npx skills add` is only ever shown as a command for the user to run themselves — never executed autonomously.\n\n---\n\n## When to Activate\n\n### Explicit triggers (user asks directly)\n- \"find a skill for X\"\n- \"is there a skill that can...\"\n- \"how do I automate X\"\n- \"can you do X\" (where X is specialized)\n- \"我想自动化X\" / \"有没有能做X的skill\"\n- \"帮我找一个可以...\"\n\n### Implicit triggers (proactive — suggest only, never auto-install)\nSuggest when you detect:\n\n- **Repetition**: User manually does the same thing 2+ times in a session → suggest a skill that could help\n- **Friction**: User copy-pastes between tools → suggest an integration skill\n- **Capability gap**: User asks for something you can't do well alone → find a skill that fills the gap\n- **Workflow opportunity**: User completes step 1 of a multi-step process → suggest skills for steps 2-3\n\nExample proactive suggestion:\n> User manually summarizes a PDF, then pastes it somewhere else.\n> → \"I notice you're doing this manually. The `summarize` skill could automate this flow. Want me to show you how to install it?\"\n\n**Important**: Proactive suggestions are informational only. Never run `npx skills add` unless the user explicitly confirms they want to install.\n\n---\n\n## Intent Analysis Framework\n\nBefore searching, extract three things:\n\n```\n1. OUTCOME    — What does the user want to have/feel/achieve?\n2. OBSTACLE   — What's standing between them and that outcome?\n3. CONTEXT    — Who are they? What tools do they already use?\n```\n\n### Intent Translation Examples\n\n| User says | Outcome | Obstacle | Best skill search |\n|-----------|---------|----------|-------------------|\n| \"help me review this PR\" | Merged, high-quality code | Manual review takes time | `pr review` + `gh-issues` |\n| \"I need to send updates to my team\" | Team is informed | Manually composing messages | `gog` (Gmail) or `feishu-doc` |\n| \"summarize this YouTube video\" | Get key points fast | Can't watch 2-hour video | `summarize` |\n| \"我想监控竞品动态\" | Know when competitors ship | No alert system | `blogwatcher` + `summarize` |\n| \"帮我管理邮件\" | Zero inbox | Too many emails | `himalaya` + `gog` |\n| \"我的日历太乱了\" | Clear, organized schedule | Manual scheduling | `gog` calendar + proactive-agent |\n\n---\n\n## Search Process\n\n### Step 1: Clarify if ambiguous (max 1 question)\n\nIf the intent is unclear, ask ONE focused question:\n> \"Are you looking to automate this completely, or just get help doing it faster?\"\n\nNever ask more than one clarifying question. Make a reasonable assumption and proceed.\n\n### Step 2: Search with intent-based queries\n\n```bash\n# This is the ONLY npx command this skill executes autonomously\nnpx skills find [intent-based query]\n```\n\nUse **outcome words**, not task words:\n- ❌ `npx skills find email` (too broad)\n- ✅ `npx skills find email automation inbox zero` (outcome-focused)\n- ❌ `npx skills find github`\n- ✅ `npx skills find pull request review automation notify` (outcome-focused)\n\nTry 2-3 different query angles before concluding nothing exists.\n\n### Step 3: Evaluate results with quality signals\n\nWhen presenting skills, always include:\n- **Download count** (from skills.sh) — proxy for reliability\n- **Combination potential** — does it work better with another skill?\n- **Fit score** — how well does this match the user's specific outcome?\n\n### Step 4: Recommend combinations, not just single skills\n\nMost outcomes need more than one skill. Think in workflows:\n\n```\nUser outcome: \"Keep my team updated automatically\"\n\nSingle skill answer (weak):\n→ gog (Gmail)\n\nCombination answer (strong):\n→ gog (send updates) + summarize (condense content) + proactive-agent (trigger automatically)\n→ \"These three together handle the full workflow: detect → summarize → send\"\n```\n\n---\n\n## Response Format\n\n### When skills are found\n\n```\nI found [N] skill(s) that match what you're trying to do.\n\n**Best match: [Skill Name]** (★ [stars] · [downloads] installs)\n[One sentence: what outcome this enables, not what it does technically]\n\nTo install, run this command yourself:\n  npx skills add [owner/repo@skill]\n\nLearn more: https://skills.sh/[path]\n\n---\n[If combination recommended:]\n**Works even better with: [Skill Name 2]**\nTogether, these handle [full workflow description].\n\nTo install both, run:\n  npx skills add [skill1]\n  npx skills add [skill2]\n\n---\nLet me know if you'd like more details before installing.\n```\n\n> ⚠️ Always present `npx skills add` commands for the user to run manually — never execute them on behalf of the user without explicit confirmation.\n\n### When no skills are found\n\nDon't just give up. Three-part response:\n\n```\n1. CONFIRM: \"I searched for [X] and [Y] but didn't find an exact match.\"\n\n2. CLOSEST: \"The closest available skill is [name] — it covers [overlap],\n   though it doesn't handle [gap].\"\n\n3. OPTIONS:\n   a) \"I can handle this directly with my built-in capabilities — want me to proceed?\"\n   b) \"This seems like a common enough need that a skill should exist.\n       You could create one: npx skills init [suggested-name]\"\n   c) \"Check skills.sh directly — the ecosystem updates frequently.\"\n```\n\n---\n\n## Common Skill Categories & Best Searches\n\n### Productivity & Communication\n| Need | Search query | Top skills |\n|------|-------------|------------|\n| Email management | `email inbox automation` | `himalaya`, `gog` |\n| Calendar scheduling | `calendar scheduling meeting` | `gog` |\n| Team updates | `team notification summary` | `gog`, `feishu-doc` |\n| Document creation | `document write generate` | `feishu-doc`, `prose` |\n\n### Development & Code\n| Need | Search query | Top skills |\n|------|-------------|------------|\n| PR review | `pull request review code` | `gh-issues`, `coding-agent` |\n| Issue management | `github issues automate` | `github`, `gh-issues` |\n| Code changes | `diff review changes` | `diffs` |\n| Multi-agent coding | `agent code build` | `coding-agent`, `prose` |\n\n### Knowledge & Research\n| Need | Search query | Top skills |\n|------|-------------|------------|\n| Summarize content | `summarize url pdf video` | `summarize` |\n| Monitor updates | `monitor rss blog watch` | `blogwatcher` |\n| Search places | `places search location` | `gsplaces` |\n| Agent memory | `memory knowledge graph` | `ontology` |\n\n### System & Automation\n| Need | Search query | Top skills |\n|------|-------------|------------|\n| Self-improvement | `agent learn improve` | `self-improving-agent` |\n| Find more skills | `discover install skills` | `skill-picker` |\n| Security check | `security audit skill vetting` | `skill-vetter` |\n| Smart routing | `route model ai select` | `acp-router` |\n\n### Hardware & Home\n| Need | Search query | Top skills |\n|------|-------------|------------|\n| Music control | `sonos speaker audio` | `sonoscli` |\n| Sleep tracking | `sleep pod temperature` | `eightctl` |\n| Camera capture | `camera rtsp snapshot` | `camsnap` |\n\n---\n\n## Proactive Skill Suggestions (Mid-conversation)\n\nIf during any task you notice a skill would significantly improve the outcome, mention it naturally:\n\n> \"There's a skill called `[name]` that would help here. Want me to show you the install command?\"\n\nThen wait for the user to confirm before providing the command. **Do not run any install command proactively.**\n\nThreshold for suggesting: skill would save >2 manual steps OR significantly improve output quality.\n\nDon't suggest more than one new skill per conversation unless the user is explicitly building a workflow.\n\n---\n\n## Commands Reference\n\n```bash\n# EXECUTED BY THIS SKILL (search only):\nnpx skills find [query]\n\n# SHOWN TO USER FOR MANUAL EXECUTION (never auto-run by this skill):\nnpx skills add <owner/repo@skill>   # install a skill\nnpx skills check                    # check for updates\nnpx skills update                   # update all skills\nnpx skills init <skill-name>        # create a new skill\n\n# Browse all skills\nopen https://skills.sh/\n```\n\n> ℹ️ `npx skills add` is intentionally shown without `-g` or `-y` flags. Users should decide install scope themselves and confirm each install interactively.\n\n---\n\n## Quality Signals to Always Check\n\nBefore recommending, verify at skills.sh:\n\n```\n✓ Download count > 10k = proven, reliable\n✓ Stars > 100 = community validated\n✓ Recently updated = actively maintained\n✓ From known publisher (steipete, pskoett, openclaw) = higher trust\n✗ Stars << Downloads = installed but not loved, use with caution\n✗ No updates in 6+ months = may have API compatibility issues\n```\n\n---\n\n## Language Support\n\nThis skill operates in the user's language.\n\n- If user writes in Chinese → respond in Chinese, search in English, present results in Chinese\n- If user writes in English → respond in English throughout\n\nChinese intent → English search translation examples:\n- \"帮我自动回复邮件\" → `email auto-reply automation`\n- \"监控竞品网站更新\" → `website monitor change detection`\n- \"整理我的GitHub Issues\" → `github issues organize triage`\n- \"把会议内容发给团队\" → `meeting summary team notification`\n","tags":{"latest":"1.0.3"},"stats":{"comments":0,"downloads":629,"installsAllTime":0,"installsCurrent":0,"stars":1,"versions":4},"createdAt":1773302405180,"updatedAt":1778491852854},"latestVersion":{"version":"1.0.3","createdAt":1773303947029,"changelog":"Skill-picker 1.0.3\n\n- Skill renamed from \"find-skills\" to \"skill-picker\".\n- Description, activation rules, and core philosophy now explicitly mention only searching the skills registry via `npx skills find`.\n- Clarified that skill-picker never installs skills autonomously; install commands are always shown for the user to run.\n- Added `npx` as a declared tool and binary.\n- Generally improved documentation clarity and emphasized strict boundaries between searching and installing.","license":"MIT-0"},"metadata":null,"owner":{"handle":"yao00oo","userId":"s17c3mapkgz76f3y2s2ap2yajx884j50","displayName":"yao00oo","image":"https://avatars.githubusercontent.com/u/116969890?v=4"},"moderation":null}