Token导航 LogoToken导航TokenDH.com
研究检索执行命令github未标认证来源可访问许可证需确认审计提醒

createcreate 开发

Agent Skill

create 用于查找、检索和筛选相关信息,适合在 Codex、Claude、Cursor、Gemini CLI 中需要根据关键词、任务场景或来源线索快速定位候选结果时使用。可结合来源仓库、安装命令和原始 README 继续核验具体用法。安装前建议确认权限范围、维护状态,以及是否会触发联网、命令执行或文件读写。

总安装

1,176

周安装

50

GitHub Stars

38

下载量

412
CodexClaudeCursorGemini CLI

安装说明

本站只整理中文说明和来源信息,不托管安装包,也不代用户安装。

GitHub

来源数

2

许可证

unknown

最后核验

2026-05-01

来源状态

来源可访问

安装方式

通过对话安装

复制提示词发给支持本地命令或 Skills 的 AI 助手,先确认命令和权限,再让它执行。

请帮我安装这个 Agent Skill:create(create 开发)
来源仓库:https://github.com/terrylica/cc-skills
仓库路径:skills/create
安装命令:
npx skills add https://github.com/terrylica/cc-skills --skill create
安装前请先检查当前环境是否支持对应 CLI,并向我确认将要执行的命令、安装目录、联网范围和文件读写权限;确认后再执行。

命令行安装

复制命令到本机终端执行。该命令会通过 npx skills 从第三方来源获取 Skill;本站只展示命令,不托管安装包,也不自动执行。

skills.shnpx skills
npx skills add https://github.com/terrylica/cc-skills --skill create

简介

create 用于查找、检索和筛选相关信息,支持基于关键词快速定位。

  • 它适用于需要根据任务场景或来源线索筛选信息的场景。
  • 可在 Codex、Claude、Cursor、Gemini CLI 中使用。
  • 建议确认权限范围和维护状态,避免触发不必要操作。
  • create 属于研究检索类 Skill,可作为该场景下的辅助能力补充。

SKILL.md

⛔ Create Plugin — STOP AND READ

DO NOT ACT ON ASSUMPTIONS. Read this file first.

This is a structured workflow command for creating a new plugin in a Claude Code marketplace.

Your FIRST and ONLY action right now: Execute the TodoWrite below.

Self-Evolving Skill: This skill improves through use. If instructions are wrong, parameters drifted, or a workaround was needed — fix this file immediately, don't defer. Only update for real, reproducible issues.

⛔ MANDATORY FIRST ACTION: TodoWrite Initialization

YOUR FIRST ACTION MUST BE the TodoWrite call below.

DO NOT:

  • ❌ Create any directories before TodoWrite
  • ❌ Read marketplace.json before TodoWrite
  • ❌ Ask questions before TodoWrite
  • ❌ Jump to any phase without completing Step 0

Step 0.1: Detect Marketplace Root

Before executing TodoWrite, verify you're in a marketplace directory:

/usr/bin/env bash << 'PREFLIGHT_EOF'
# Check for marketplace.json in cwd
if [ -f ".claude-plugin/marketplace.json" ]; then
  echo "✅ Marketplace detected: $(jq -r .name .claude-plugin/marketplace.json)"
else
  echo "❌ Not a marketplace directory. Run from a marketplace root."
  exit 1
fi
PREFLIGHT_EOF

Step 0.2: Execute MANDATORY TodoWrite

Execute TodoWrite NOW with this template:

TodoWrite with todos:

- "[Plugin] Phase 0: Detect marketplace root" | in_progress
- "[Plugin] Phase 0: Interactive prompts (name, category, components)" | pending
- "[Plugin] Phase 0: Confirm plugin doesn't exist" | pending
- "[Plugin] Phase 1: Skill → plugin-structure (scaffold)" | pending
- "[Plugin] Phase 1: Create plugin directory + plugin.json" | pending
- "[Plugin] Phase 1: Skill → implement-plan-preflight (ADR)" | pending
- "[Plugin] Phase 2: Skill → skill-architecture (if has-skills)" | pending
- "[Plugin] Phase 2: Skill → hook-development (if has-hooks)" | pending
- "[Plugin] Phase 2: Skill → command-development (if has-commands)" | pending
- "[Plugin] Phase 2: Skill → agent-development (if has-agents)" | pending
- "[Plugin] Phase 2: Agent → skill-reviewer (if skills created)" | pending
- "[Plugin] Phase 3: Add to marketplace.json" | pending
- "[Plugin] Phase 3: Run validate-plugins.mjs" | pending
- "[Plugin] Phase 3: Skill → code-hardcode-audit" | pending
- "[Plugin] Phase 3: Agent → plugin-validator" | pending
- "[Plugin] Phase 4: Git commit (conventional format)" | pending
- "[Plugin] Phase 4: Push to remote" | pending
- "[Plugin] Phase 4: Skill → semantic-release" | pending

After TodoWrite completes, proceed to Phase 0 section below.


Quick Reference

Skills Invoked (Optimized Sequence)

OrderSkillPhasePurposeInvocation
1plugin-structure1Directory & manifestSkill(plugin-dev:plugin-structure)
2implement-plan-preflight1ADR + Design Spec + DiagramsSkill(itp:implement-plan-preflight)
3skill-architecture2Create skills (if has-skills)Skill(plugin-dev:skill-architecture)
4hook-development2Create hooks (if has-hooks)Skill(plugin-dev:hook-development)
5command-development2Create commands (if has-commands)Skill(plugin-dev:command-development)
6agent-development2Create agents (if has-agents)Skill(plugin-dev:agent-development)
7code-hardcode-audit3Quality auditSkill(itp:code-hardcode-audit)
8plugin-validator3Silent failure auditSkill(plugin-dev:plugin-validator)
9run-full-release4Run repo's mise release pipelineSkill(mise:run-full-release)

Skills EXCLUDED (Redundant)

SkillReason Excluded
plugin-dev:skill-developmentUse skill-architecture instead (3x more comprehensive)
plugin-dev:plugin-settingsMerged into hook-development
itp:adr-graph-easy-architectInvoked BY implement-plan-preflight (not separately)

Agents Spawned

PhaseAgentPurposeInvocation
2skill-reviewerReview skill qualityTask(plugin-dev:skill-reviewer)
3plugin-validatorValidate final structureTask(plugin-dev:plugin-validator)

File Locations

ArtifactPathNotes
Plugin Directoryplugins/{name}/Main plugin folder
Plugin Manifestplugins/{name}/plugin.jsonRequired manifest
Plugin READMEplugins/{name}/README.mdDocumentation
Marketplace JSON.claude-plugin/marketplace.jsonMust add plugin entry
ADRdocs/adr/YYYY-MM-DD-{name}.mdCreated by preflight skill
Design Specdocs/design/YYYY-MM-DD-{name}/spec.mdCreated by preflight skill

Phase 0: Discovery & Validation

Detect marketplace root, gather plugin name/category/components via interactive prompts, confirm the plugin does not already exist.

Detailed steps: Phase 0 Reference

Phase 0 Gate

  • Marketplace root detected (.claude-plugin/marketplace.json exists)
  • Plugin name collected (kebab-case, no spaces)
  • Category selected
  • Components selected (skills/hooks/commands/agents)
  • Plugin directory does NOT exist
  • Plugin NOT in marketplace.json

Phase 1: Scaffold Plugin

  1. Invoke Skill(plugin-dev:plugin-structure) -- directory & manifest patterns
  2. Create plugin directory with component subdirs based on user selections
  3. Generate plugin.json using marketplace version
  4. Invoke Skill(itp:implement-plan-preflight) -- creates ADR + Design Spec + diagrams

Detailed steps: Phase 1 Reference

Phase 1 Gate

  • Plugin directory exists: plugins/$PLUGIN_NAME/
  • plugin.json created with marketplace version
  • ADR exists: docs/adr/$ADR_ID.md
  • Design spec exists: docs/design/$ADR_ID/spec.md
  • Both diagrams in ADR (Before/After + Architecture)

Phase 2: Component Creation (Conditional)

Execute ONLY the skills for components the user selected:

ComponentSkill InvocationThen
SkillsSkill(plugin-dev:skill-architecture)Spawn Task(plugin-dev:skill-reviewer)
HooksSkill(plugin-dev:hook-development)--
CommandsSkill(plugin-dev:command-development)--
AgentsSkill(plugin-dev:agent-development)--

Detailed steps: Phase 2 Reference

Phase 2 Gate

  • All selected components created
  • If skills: skill-reviewer agent completed review
  • Files follow plugin-dev patterns

Phase 3: Registration & Validation

  1. Add plugin to marketplace.json (include hooks field if hooks exist)
  2. Run node scripts/validate-plugins.mjs -- expect all-pass
  3. Invoke Skill(itp:code-hardcode-audit) -- quality audit
  4. Run silent failure audit on all hook entry points
  5. Spawn Task(plugin-dev:plugin-validator) -- structural validation

Detailed steps: Phase 3 Reference

Phase 3 Gate

  • Plugin added to marketplace.json
  • validate-plugins.mjs passes
  • code-hardcode-audit passes
  • silent-failure-audit passes (no errors)
  • plugin-validator agent approves

Phase 4: Commit & Release

  1. Stage changes: plugin dir, marketplace.json, ADR, design spec
  2. Conventional commit: feat($PLUGIN_NAME): add plugin for [brief description]
  3. Push to remote
  4. Invoke Skill(mise:run-full-release) -- run repo's mise release pipeline for tag, changelog, GitHub release

Detailed steps: Phase 4 Reference

Phase 4 Success Criteria

  • All changes committed with conventional commit
  • Pushed to remote
  • semantic-release completed
  • New version tag created
  • GitHub release published

Troubleshooting

See Troubleshooting Reference for common issues and fixes.

Post-Execution Reflection

After this skill completes, reflect before closing the task:

  1. Locate yourself. — Find this SKILL.md's canonical path (Glob for this skill's name) before editing. All corrections target THIS file and its sibling references/ — never other documentation.
  2. What failed? — Fix the instruction that caused it. If it could recur, add it as an anti-pattern.
  3. What worked better than expected? — Promote it to recommended practice. Document why.
  4. What drifted? — Any script, reference, or external dependency that no longer matches reality gets fixed now.
  5. Log it. — Every change gets an evolution-log entry with trigger, fix, and evidence.

Do NOT defer. The next invocation inherits whatever you leave behind.



适合场景

01

用户想查找某类 Agent Skill 时

02

需要根据任务场景推荐可安装能力包时

03

需要对比不同来源的安装命令和来源信息时

能力概览

能力 1

按任务关键词查找相关 Skills

能力 2

展示可复制的安装命令

能力 3

保留来源站点、仓库和原始说明,方便继续核验

能力 4

展示第三方安全扫描或审计结果

安装后应在对应宿主中按原始 README 的触发条件使用;具体调用方式请以来源页面和 README 为准。

平台分布

Codex

34.67%
按下载量换算143

Claude

30.4%
按下载量换算125

Cursor

18.6%
按下载量换算77

Gemini CLI

9.96%
按下载量换算41

安全审计

Gen Agent Trust Hub

通过

Socket

可疑

Snyk

通过

权限和风险

执行命令

安装流程涉及命令执行,可能通过 npx skills add https://github.com/terrylica/cc-skills --skill create 联网下载 Skill 或依赖。用户安装前应确认命令来源、仓库内容和执行环境。

安装前确认

本站仅展示第三方公开信息,不托管安装包,不提供自动安装或运行环境。安装前应自行审查源码、依赖和命令行为。来源安全扫描存在 warning/failed 结果,不能写成本站确认安全。当前只有一个来源,正式发布前建议补源仓库或其他目录站核验。

来源信息

继续浏览同类 Skills