skillshare
Manages and syncs AI CLI skills and agents across 50+ tools from a single source. Use this skill whenever the user mentions "skillshare", runs skillshare commands, manages skills or agents (install, update, uninstall, sync, commit, audit, analyze, check, diff, search), or troubleshoots skill/agent configuration (orphaned symlinks, broken targets, sync issues). Covers both global (~/.config/skillshare/) and project (.skillshare/ or skillshare/) modes. Also use when: adding new AI tool targets (Claude, Cursor, Windsurf, etc.), setting target include/exclude filters or copy vs symlink mode, using backup/restore or trash recovery, piping skillshare output to scripts (--json), setting up CI/CD audit pipelines, building/sharing skill hubs (hub index, hub add), or working with agents (single .md files synced to agent-capable targets like Claude, Cursor, Augment, OpenCode) via positional `agents` filter or `--kind agent`, plus `.agentignore` and `enable`/`disable` for per-agent toggles.
获取
https://deepseekmodel.com/api/download.php?id=runkids-skillshare-skills-skillshare-skill-md&format=skill
name skillshare description Manages and syncs AI CLI skills and agents across 50+ tools from a single source. Use this skill whenever the user mentions "skillshare", runs skillshare commands, manages skills or agents (install, update, uninstall, sync, commit, audit, analyze, check, diff, search), or troubleshoots skill/agent configuration (orphaned symlinks, broken targets, sync issues). Covers both global (~/.config/skillshare/) and project (.skillshare/ or skillshare/) modes. Also use when: adding new AI tool targets (Claude, Cursor, Windsurf, etc.), setting target include/exclude filters or copy vs symlink mode, using backup/restore or trash recovery, piping skillshare output to scripts (--json), setting up CI/CD audit pipelines, building/sharing skill hubs (hub index, hub add), or working with agents (single .md files synced to agent-capable targets like Claude, Cursor, Augment, OpenCode) via positional `agents` filter or `--kind agent`, plus `.agentignore` and `enable`/`disable` for per-agent toggles. argument-hint [command] [target] [--json] [--dry-run] [-p|-g] metadata {"version":"v0.20.27"} Skillshare CLI Global: ~/.config/skillshare/skills/ → all AI CLIs. Project: .skillshare/skills/ → repo-local. Auto-detects project mode from .skillshare/config.yaml or skillshare/config.yaml (hidden wins when both exist). Force with -p or -g . Recipes Getting Started skillshare init --no-copy --all-targets --git --skill # Fresh global setup skillshare init -p --targets "claude,cursor" # Fresh project setup skillshare init -p --visible # Use skillshare/ instead of .skillshare/ skillshare init --copy-from claude --all-targets --git # Import from existing CLI skillshare init --discover -- select "windsurf" # Add new AI tool later Installing Skills skillshare install user/repo -s pdf,commit # Select specific skills skillshare install user/repo --all # Install everything skillshare install user/repo --into frontend # Place in subdirectory skillshare install gitlab.com/team/repo # Any Git host skillshare install user/repo --track # Enable `update` later skillshare install user/repo -b develop --all # Install from branch skillshare install user/repo --track -b develop # Track specific branch skillshare install user/repo -s pdf -p # Install to project skillshare install # Rehydrate remote skills/tracked repos skillshare sync # Always sync after install Extras (Rules, Commands, Prompts) skillshare extras init rules --target ~/.claude/rules --target ~/.cursor/rules skillshare extras init commands --target ~/.claude/commands --mode copy skillshare extras init rules --target ~/.claude/rules -- source ~/shared/rules # custom source (global only) skillshare extras init rules --target ~/.cursor/rules --force # overwrite existing skillshare extras init # Interactive TUI wizard (incl. source step) skillshare extras source # Show current extras_source skillshare extras source ~/shared/extras # Set global extras_source skillshare extras list # Show status per target skillshare extras list --json # JSON with source_type field skillshare extras collect rules # Pull local files into source skillshare extras remove rules # Remove from config (source preserved) skillshare extras init agents --target ~/.claude/agents --flatten # Flatten subdirs into root skillshare extras rules --mode copy # Change sync mode of a target skillshare extras agents --flatten # Enable flatten on existing target skillshare extras rules --add-target ~/.cursor/rules # Add a target to an existing extra skillshare extras rules --remove-target ~/.cursor/rules --prune # Remove a target (--prune deletes synced files) skillshare sync extras # Sync all extras to targets skillshare sync extras --dry-run --force # Preview / overwrite conflicts skillshare sync --all # Sync skills + extras together Project extras always read from <sources.extras>/<name> (default .skillshare/extras/<name> ); per-extra source and extras source are global-only. For project agents, prefer target agents: unless you need extras-only flatten/extension. A target can set an extension: field in config.yaml to transform each source file during sync (e.g. markdown → TOML for Gemini/Codex); implies copy mode. See extras.md for details. Extensions (Content Transformation) # Extension is a directory at .skillshare/extensions/<name>/ # extension.yaml controls: run command, output_ext, description # Official extensions: https://github.com/runkids/skillshare/tree/main/extensions extension.yaml format: run: [ "node" , "convert.js" ] # or ["python3", "convert.py"] etc. output_ext: toml # renames output file extension description: "MD → Codex TOML" Add extension to an extras target in config.yaml: extras: - name: agents targets: - path: .claude/agents # plain copy/merge - path: .codex/agents extension: codex-agents # transform + rename to .toml Caveats: extension: is only supported on extras targets — native agents targets ignore it Implies copy mode; merge/symlink won't apply Claude Code sets NODE_OPTIONS that break node extensions — use env -u NODE_OPTIONS node convert.js in run Files missing required frontmatter fields fail validation and are skipped Creating & Discovering Skills skillshare new my-skill # Create with interactive pattern selection skillshare new my-skill -P reviewer # Use reviewer pattern directly skillshare search "react testing" # Search GitHub for skills skillshare collect # Pull target-local changes back to source Removing Skills skillshare uninstall my-skill # Remove one (moves to trash) skillshare uninstall skill-a skill-b # Remove multiple skillshare uninstall -G frontend # Remove entire group skillshare sync # Always sync after uninstall Enable / Disable Skills skillshare disable draft-* # Hide from sync (adds to .skillignore) skillshare enable draft-* # Restore (removes from .skillignore) skillshare disable my-skill -p # Project mode skillshare disable my-skill --dry-run # Preview # TUI: press E in `skillshare list` to toggle skillshare sync # Always sync after toggle Team / Organization # Creator: init project (see Getting Started) → add skills → commit .skillshare/ skillshare install -p && skillshare sync # Member: clone → install → sync skillshare install github.com/team/repo --track -p # Track shared repo skillshare commit -m "Update skill" # Local checkpoint, no push skillshare push # Cross-machine: push on A skillshare pull # Cross-machine: pull on B Skill Hubs skillshare hub add https://example.com/hub.json # Save a hub source skillshare hub add https://example.com/hub.json --label my-hub # With custom label skillshare hub add git@ghe.corp.com:team/skills.git --label ghe # SSH/private/GHE hub source skillshare hub list # List saved hubs skillshare hub default my-hub # Set default hub skillshare hub remove my-hub # Remove a hub skillshare hub index -- source ~/.config/skillshare/skills/ --full --audit # Build hub index Controlling Where Skills Go # SKILL.md frontmatter: metadata.targets: [claude] → only syncs to Claude skillshare target claude --add-include "team-*" # glob filter skillshare target claude --add-agent-include "team-*" # agent glob filter skillshare target claude --add-exclude "_legacy*" # exclude pattern skillshare target claude --agent-mode copy # agents copy mode skillshare target codex --mode copy && skillshare sync --force # copy mode # .skillignore — hide skills/dirs from discovery (gitignore syntax) # Root-level: <source>/.skillignore (affects all commands) # Repo-level: <source>/_repo/.skillignore (scoped to that repo) # .skillignore.local — local override (not committed), negation overrides base See targets.md for details. Updates & Maintenance skillshare check # See what has updates skillshare update my-skill && skillshare sync # Update one skillshare update --all && skillshare sync # Update all skillshare update --all --diff # Show what changed Scripting & CI/CD skillshare status --json # Full status as JSON skillshare check --json # Update status as JSON skillshare sync --json # Sync results as JSON skillshare diff --json # Diff results as JSON skillshare install user/repo --json # Install result as JSON (implies --force --all) skillshare update --all --json # Update results as JSON skillshare uninstall my-skill --json # Global JSON; dirty tracked repos still require --force skillshare collect claude --json # Collect result as JSON; existing items still require --force skillshare target list --json # Target list as JSON skillshare list --json # Skill list as JSON skillshare search react --json # Search results as JSON skillshare audit --format json # Audit results as JSON skillshare doctor --json # Health check as JSON (exit 1 on errors) Recovery & Troubleshooting skillshare trash restore <name> && skillshare sync # Undo delete skillshare sync # Skill missing? Re-sync skillshare doctor && skillshare status # Diagnose issues skillshare install user/repo --force # Override audit block skillshare install user/repo --skip-audit # Bypass scan entirely See TROUBLESHOOTING.md for more. Quick Lookup Commands Project? --json ? status , diff , list , doctor ✓ (auto) ✓ sync , collect ✓ (auto) ✓ install , uninstall , update , check , search , new ✓ ( -p ) ✓ (except new) target , audit , analyze , trash , log , hub ✓ ( -p ) ✓ (target list, audit, analyze, log) extras init/list/remove/collect/source (+ --mode / --add-target / --remove-target flags on extras <name> ) ✓ ( -p , except source) ✓ (list, mode, targets) enable , disable ✓ (auto) ✗ commit , push , pull , backup , restore ✗ ✗ tui , upgrade ✗ ✗ ui ✓ ( -p ) ✗ AI Caller Rules Non-interactive — AI cannot answer prompts. Use --force , --all , -s , --targets , --no-copy , --all-targets , --yes . Sync after mutations — install , uninstall , update , collect , target all need sync . Audit — install auto-scans; CRITICAL blocks. --force to override, --skip-audit to bypass. Detects hardcoded secrets (API keys, tokens, private keys). Uninstall safely — moves to trash (7 days). trash restore <name> to undo. NEVER rm -rf symlinks. Output — --json for structured data (12 commands support it, see Quick Lookup). --no-tui for plain text on TUI commands ( list , log , audit , analyze , diff , trash list , backup list , target list ). tui off disables TUI globally. --dry-run to preview. Node extensions in Claude Code — NODE_OPTIONS is preloaded by the Claude Code harness and causes node to crash when running extensions. Always use run: ["env", "-u", "NODE_OPTIONS", "node", "convert.js"] in extension.yaml when writing extensions that use Node.js. References Topic File Init flags init.md Sync/collect/commit/push/pull sync.md Install/update/uninstall/new install.md Status/diff/list/search/check status.md Security audit audit.md Trash trash.md Operation log log.md Targets targets.md Extras (rules/commands/prompts) extras.md Backup/restore backup.md Troubleshooting TROUBLESHOOTING.md
该技能未提供触发词。
| 字段 | 说明 |
|---|---|
| format | 格式标识(skill/v1) |
| skill_id | 技能唯一 ID |
| name | 技能名称 |
| version | 版本号 |
| description | 技能描述 |
| category | 所属分类(数组) |
| trigger_words | 触发词列表 |
| tags | 标签列表 |
| source | 来源标识 |
| source_url | 来源链接(本页地址) |
| exported_at | 导出时间(每次下载生成) |
| system_prompt | 系统提示词正文 |
| model_config | 模型参数:provider / model / temperature / max_tokens / top_p |
| examples | 示例 |
| install_guide | 各平台导入说明(Coze / Dify / Claude / 自定义框架) |