Token导航 LogoToken导航TokenDH.com
研究检索只读github未标认证来源可访问许可证需确认审计通过

artifact-lifecycle工件生命周期

Agent Skill

artifact-lifecycle 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

1,360

周安装

55

GitHub Stars

25

下载量

427
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:artifact-lifecycle(工件生命周期)
来源仓库:https://github.com/oimiragieo/agent-studio
仓库路径:skills/artifact-lifecycle
安装命令:
npx skills add https://github.com/oimiragieo/agent-studio --skill artifact-lifecycle
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/oimiragieo/agent-studio --skill artifact-lifecycle

简介

artifact-lifecycle 提供统一的工件生命周期管理接口,简化创建、更新和废弃流程。

  • 它封装了完整的工件管理流程,支持技能、代理、钩子等类型的管理。
  • 可通过简单调用 Skill({skill: 'artifact-lifecycle'}) 启动操作。
  • 使用前应确认目标工件路径和类型以避免错误操作。
  • 适用宿主包括 Codex、Claude、Cursor、Gemini CLI,接入前应确认版本、权限和运行环境要求。

SKILL.md

Artifact Lifecycle Skill

Convenience wrapper for the unified artifact lifecycle workflow.

Overview

This skill provides a simplified interface to the comprehensive artifact lifecycle management workflow at .claude/workflows/core/skill-lifecycle.md.

When to Use

Use this skill when you need to:

  • Create a new artifact (skill, agent, hook, workflow, template, schema)
  • Update an existing artifact
  • Deprecate an artifact with migration guidance
  • Check if an artifact already exists

Quick Start

// Invoke this skill
Skill({ skill: 'artifact-lifecycle' });

Workflow Reference

Full Workflow: .claude/workflows/core/skill-lifecycle.md

Phases

Phase 1: Discovery

Check if artifact exists, compare versions.

Agent: architect Output: Discovery report with recommendations

Phase 2: Decision

Determine action: CREATE, UPDATE, or DEPRECATE.

Agent: planner Output: Action plan with tasks

Phase 3: Action

Execute the determined action.

Agent: developer + appropriate creator skill Tools:

  • Skill({skill: "skill-creator"}) for skills
  • Skill({skill: "agent-creator"}) for agents
  • Skill({skill: "hook-creator"}) for hooks
  • Skill({skill: "workflow-creator"}) for workflows

Phase 4: Integration

Update registries, catalogs, and CLAUDE.md.

Agent: developer Updates:

  • creator-registry.json
  • skill-catalog.md (for skills)
  • CLAUDE.md Section 3 (for agents) or Section 8.5/8.6 (for skills/workflows)

Phase 5: Validation

Test integration and verify references.

Agent: qa Checks:

  • Artifact invocable
  • References valid
  • No broken cross-references

Usage Examples

Create New Skill

User: "Create a skill for Kubernetes deployment"

Router spawns with:
Skill({ skill: "artifact-lifecycle" })

Workflow determines: CREATE mode
Invokes: skill-creator
Updates: registry, catalog, CLAUDE.md
Validates: skill invocable

Update Existing Agent

User: "Update the devops agent to support Terraform Cloud"

Router spawns with:
Skill({ skill: "artifact-lifecycle" })

Workflow determines: UPDATE mode
Version: 1.0.0 → 1.1.0
Updates: agent file, CHANGELOG
Validates: agent referenced correctly

Deprecate Workflow

User: "Deprecate the old deployment workflow"

Router spawns with:
Skill({ skill: "artifact-lifecycle" })

Workflow determines: DEPRECATE mode
Adds: deprecation notice, migration guide
Updates: CLAUDE.md with replacement reference
Validates: no broken references

Configuration

ParameterValuesDefault
artifact_typeskill, agent, hook, workflow, template, schemaauto-detect
operationcreate, update, deprecate, integrateauto-detect
version_bumpmajor, minor, patchminor

Iron Laws

  1. ALWAYS use type-specific creator skills — never treat artifact-lifecycle as a shortcut that bypasses research-synthesis, companion-check, or validation; each phase exists for a reason.
  2. ALWAYS determine operation mode before starting action (CREATE/UPDATE/DEPRECATE) — never begin Phase 3 without a confirmed Phase 2 decision; wrong operation = wasted or dangerous work.
  3. NEVER skip Phase 5 validation — artifacts must be verified as invocable before the lifecycle task is marked complete; broken artifacts delivered as complete are worse than none.
  4. ALWAYS update catalog and CLAUDE.md in Phase 4 — artifacts without catalog/routing entries are invisible to agents and the router; integration without registration is incomplete.
  5. NEVER use artifact-lifecycle for simple single-artifact updates — use type-specific updaters (skill-updater, agent-updater, workflow-updater); lifecycle is for multi-phase orchestration only.

Anti-Patterns

Anti-PatternWhy It FailsCorrect Approach
Skipping research-synthesis before CREATERisk of creating a duplicate or outdated artifactAlways run research-synthesis first
Starting action before Phase 2 decisionWrong operation applied; CREATE when UPDATE neededComplete Phase 2 before Phase 3
Skipping Phase 5 validationBroken artifacts delivered as completeAlways run invocability check
No catalog/registry update in Phase 4Artifact is invisible to agents and routerAlways update catalog, registry, and CLAUDE.md
Using artifact-lifecycle for simple version bumpsHeavyweight process for lightweight workUse skill-updater/agent-updater for UPDATE mode
Not specifying artifact_typeWrong creator skill invoked silentlyAlways specify or verify artifact_type in Phase 1

Memory Protocol

  1. Read .claude/context/memory/learnings.md before starting
  2. Record decisions to .claude/context/memory/decisions.md
  3. Record issues to .claude/context/memory/issues.md

Related Skills

  • skill-creator - Direct skill creation
  • agent-creator - Direct agent creation
  • workflow-creator - Direct workflow creation
  • codebase-integration - External artifact integration

Related Workflows

  • .claude/workflows/core/skill-lifecycle.md - Full lifecycle workflow
  • .claude/workflows/core/external-integration.md - External artifact integration
  • .claude/workflows/core/router-decision.md - Router decision flow

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

38.09%
按下载量换算163

Claude

29.66%
按下载量换算127

Cursor

19.59%
按下载量换算84

Gemini CLI

9.63%
按下载量换算41

安全审计

Gen Agent Trust Hub

通过

Socket

通过

Snyk

通过

权限和风险

只读

该 Skill 主要提供规则、说明或参考内容,本身偏只读;真正读写文件、联网或执行命令仍取决于宿主 Agent 的任务。

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。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