Token导航 LogoToken导航TokenDH.com
研究检索external-servicegithub未标认证来源可访问clear审计未展示

session-start-routine会话开始例程

Agent Skill

session-start-routine 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

441

周安装

18

GitHub Stars

公开资料未说明

下载量

143
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

MIT

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:session-start-routine(会话开始例程)
来源仓库:https://github.com/vamseeachanta/workspace-hub
仓库路径:skills/session-start-routine
安装命令:
npx skills add vamseeachanta/workspace-hub --skill "session-start-routine"
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。

AgentSkills.tonpx skills
npx skills add vamseeachanta/workspace-hub --skill "session-start-routine"

简介

session-start-routine 用于查找、检索和筛选相关信息,适合在 Codex、Claude、Cursor、Gemini CLI 中需要根据关键词快速定位候选结果时使用。

  • 适用于会话开始例程相关研究检索任务,可结合来源仓库进一步核验具体用法。
  • 通过 npx skills add vamseeachanta/workspace-hub --skill "session-start-routine" 命令安装。
  • 安装前建议确认权限范围和维护状态,注意可能触发联网或文件读写操作。
  • 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

Session Start Routine Skill

Version: 2.0.0 Created: 2025-12-30 Last Updated: 2026-01-02

Overview

This meta-skill defines the mandatory work session initialization routine. At the start of every conversation, Claude should execute this routine to maintain and improve the skills library.

Quick Start

  1. Trigger - Say "session start" or /session-start-routine
  2. Health check - Claude scans skill locations
  3. Review report - See skills health summary
  4. Act on findings - Update stale skills if needed
# Quick health check command
find ~/.claude/skills -name "SKILL.md" -mtime +30 -exec echo "Stale: {}" \;

# Count skills
find ~/.claude/skills -name "SKILL.md" | wc -l

When to Use

  • MANDATORY at every session start
  • First message in a new conversation
  • When explicitly requested with /session-start-routine
  • After extended breaks between sessions
  • When noticing skill-related issues

When to Trigger

MANDATORY at every session start:

  • First message in a new conversation
  • When explicitly requested with /session-start-routine

Session Routine Steps

Step 1: Skills Health Check

1. Scan both skill locations:
   - User-level: ~/.claude/skills/
   - Project-level: .claude/skills/

2. For each skill:
   - Verify SKILL.md exists and is valid
   - Check for version history section
   - Identify last update date
   - Flag skills not updated in 30+ days

3. Dependency Verification:
   - Check if tools/libraries referenced in skills still exist
   - Verify MCP servers mentioned are still available
   - Test that code examples are still valid
   - Flag skills with broken dependencies

Step 1b: Integration Check

1. MCP Server Compatibility:
   - List skills that reference MCP servers
   - Verify those servers exist in ~/.claude.json or .mcp.json
   - Check for deprecated MCP tool names

2. Agent Compatibility:
   - Verify skills work with current Claude Code agents
   - Check for deprecated agent patterns
   - Update agent references if needed

Step 1c: Cross-Skill Dependencies

1. Build dependency map:
   - Extract "Related Skills" sections
   - Identify which skills reference others
   - Detect orphaned or circular dependencies

2. Dependency Graph:
   skill-creator
     +-- sparc-workflow
     +-- mcp-builder
   rag-system-builder
     +-- knowledge-base-builder
     +-- semantic-search-setup
     +-- pdf-text-extractor

Step 2: Research & Improvement Identification

For each skill category, research:
1. Official documentation updates
   - Anthropic Claude documentation
   - Tool/library changelogs

2. Community best practices
   - GitHub trending repositories
   - Stack Overflow discussions

3. Industry trends
   - New tools/libraries
   - Methodology updates
   - Security advisories

Step 3: Skill Updates

For skills needing updates:
1. Research latest best practices online
2. Identify specific improvements:
   - New features to document
   - Deprecated approaches to remove
   - Better code examples
   - Performance optimizations

3. Apply updates with version comment:
   ---
   ## Version History
   - **X.Y.Z** (YYYY-MM-DD): [Changes made]
   ---

Step 4: New Skill Identification

Evaluate need for new skills based on:

1. GAPS: Missing capabilities discovered during work
   - "I needed to do X but no skill existed"
   - Repeated manual processes

2. PATTERNS: Common workflows across sessions
   - Frequently used tool combinations
   - Repeated prompt patterns

3. TRENDS: Industry/technology developments
   - New Claude Code features
   - New MCP servers
   - Popular libraries/tools
   - AI/ML advancements

Step 5: Documentation Update

1. Update skills/README.md with:
   - New skills added
   - Skills updated
   - Skills deprecated

2. Commit changes with descriptive message

Session Report Template

At the end of routine, provide summary:

## Session Start Report - [DATE]

### Skills Health
- Total Skills: [N]
- User-Level: [N] | Project-Level: [N]
- Updated This Session: [N]
- Flagged for Review: [N]

### Improvements Made
| Skill | Version | Changes |
|-------|---------|---------|
| skill-name | X.Y.Z | Brief description |

### New Skills Identified
| Proposed Skill | Category | Rationale |
|----------------|----------|-----------|
| skill-name | category | Why needed |

### Research Findings
- [Key finding 1]
- [Key finding 2]

### Next Session Priorities
1. [Priority 1]
2. [Priority 2]

Execution Checklist

Before routine:

  • Session context established
  • Access to skill directories confirmed
  • Network access for research (optional)

During routine:

  • User-level skills scanned
  • Project-level skills scanned
  • Stale skills identified (30+ days)
  • Broken dependencies flagged
  • Cross-skill dependencies mapped

After routine:

  • Session report generated
  • Priority actions identified
  • Updates committed (if any)
  • README updated (if changes)

Skill Update Guidelines

Version Numbering

MAJOR.MINOR.PATCH

MAJOR: Breaking changes, complete rewrites
MINOR: New features, significant improvements
PATCH: Bug fixes, small updates, typo fixes

Version Comment Format

Add at end of SKILL.md:

---

## Version History

- **1.2.0** (2025-12-30): Added new section for X, updated examples
- **1.1.1** (2025-12-15): Fixed typo in command example
- **1.1.0** (2025-12-01): Added integration with tool Y
- **1.0.0** (2025-11-15): Initial release

Error Handling

Common Errors

Error: Skill directory not accessible

  • Cause: Permissions or path issues
  • Solution: Verify ~/.claude/skills/ exists and is readable

Error: SKILL.md parsing fails

  • Cause: Invalid YAML frontmatter or markdown
  • Solution: Validate frontmatter syntax, check for unclosed code blocks

Error: MCP server not found

  • Cause: Referenced MCP server not configured
  • Solution: Update skill to remove reference or add server to config

Error: Circular dependency detected

  • Cause: Skills reference each other in loop
  • Solution: Review related_skills, break the cycle

Error: Too many stale skills

  • Cause: Routine not run regularly
  • Solution: Prioritize most-used skills, update in batches

Research Sources

Official Documentation

Community Resources

Technology Updates

  • Python/Node.js changelogs
  • Library release notes
  • Security advisories (CVE databases)

Integration Points

With Other Skills

  • Uses skill-creator patterns for new skills
  • Follows compliance-check standards
  • Updates workspace-hub skills as needed

With Workflows

  • Can trigger before SPARC workflow
  • Integrates with git operations for commits
  • Works with Agent OS task execution

Automation Hooks

Pre-Session Check

# Quick health check script
find ~/.claude/skills -name "SKILL.md" -mtime +30 -exec echo "Stale: {}" \;

Post-Update Commit

# Commit pattern for skill updates
git add ~/.claude/skills/
git commit -m "chore(skills): session update - [YYYY-MM-DD]

- Updated: [skill1], [skill2]
- Added: [new-skill]
- Research: [topic]"

Usage Analytics

Track skill usage to prioritize improvements:

1. Usage Tracking:
   - Note which skills are triggered during sessions
   - Track frequency of skill invocations
   - Identify most/least used skills

2. Prioritization Matrix:
   HIGH PRIORITY (Frequently used, needs update):
   - mcp-builder (used 50+/month, last update 30+ days)

   MEDIUM PRIORITY (Used often, recently updated):
   - rag-system-builder (used 30+/month, updated recently)

   LOW PRIORITY (Rarely used):
   - algorithmic-art (used 2/month)

3. Usage Log Format:
   ~/.claude/skills/.usage-log.json
   {
     "2025-12-30": {
       "mcp-builder": 3,
       "rag-system-builder": 2,
       "sparc-workflow": 5
     }
   }

Deprecation Review

Flag outdated skills for removal or major update:

1. Deprecation Criteria:
   - Not updated in 90+ days
   - References deprecated tools/APIs
   - Superseded by better skill
   - Zero usage in 60+ days

2. Deprecation Process:
   a. Add DEPRECATED prefix to description
   b. Add deprecation notice at top of SKILL.md
   c. Point to replacement skill
   d. Remove after 30-day grace period

3. Deprecation Notice Template:
   ---
   name: old-skill-name
   description: DEPRECATED - Use new-skill-name instead. [Original description]
   deprecated: true
   deprecated_date: 2025-12-30
   replacement: new-skill-name
   ---

   > **DEPRECATED**: This skill is deprecated. Use [new-skill-name](../new-skill-name/SKILL.md) instead.
   > Will be removed after 2025-01-30.

Metrics

Track over time:

MetricTargetHow to Measure
Skills updated per session1-3Count of version bumps
New skills added per month2-5New SKILL.md files
Stale skill ratio<20%Skills 30+ days old / total
Trigger accuracy>90%Correct activations
Integration health100%Working MCP/agent refs
Deprecation candidates<5%Unused 60+ days

Best Practices

  1. Consistency: Run at every session start
  2. Brevity: Keep updates focused, don't over-engineer
  3. Documentation: Always add version comments
  4. Research First: Check online before updating
  5. Prioritize: Focus on most-used skills first
  6. Test: Verify updated skills still work correctly

Related Skills

Note: Skills may be at user-level (~/.claude/skills/) or workspace-level (workspace-hub/.claude/skills/)
SkillPurposeLocation
skill-creatorCreate new skillsworkspace-hub/.claude/skills/builders/skill-creator/
context-managementManage context limits.claude/skills/context-management/
dev-workflowDevelopment workflow.claude/skills/workflows/dev-workflow/
codex-reviewCross-review with Codex.claude/skills/workflows/codex-review/

Version History

  • 2.0.1 (2026-01-20): Fixed related skills links - replaced broken relative paths with table format showing actual locations
  • 2.0.0 (2026-01-02): Upgraded to v2 template - added Quick Start, When to Use, Execution Checklist, Error Handling, enhanced Metrics sections; enhanced frontmatter with version, category, related_skills
  • 1.0.0 (2025-12-30): Initial release - session start routine skill

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

04

需要参考平台分布和安装热度时

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

补充不同宿主或平台的使用分布数据

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

Claude Code

33.55%
按下载量换算48

windsurf

22.66%
按下载量换算32

trae

17.02%
按下载量换算24

OpenCode

12.05%
按下载量换算17

Cursor

8.58%
按下载量换算12

Codex

3.31%
按下载量换算5

安全审计

暂无安全审计结果可展示。

权限和风险

external-service

该 Skill 可能调用第三方服务、云服务或外部模型 API,使用前需要确认账号、额度、数据发送范围和服务条款。

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