AgentSkillsCN

session-status

展示当前会话中正在使用的Claude Code原语。

SKILL.md
--- frontmatter
name: session-status
description: Show what Claude Code primitives are active in this session

Session Status Check

When invoked, analyze and report the current session state.

Instructions

  1. Identify the project from the working directory
  2. List active primitives by checking what's loaded
  3. Show auto-invoke rules from CLAUDE.md
  4. Flag any issues (missing skills, broken references)

Output Format

code
┌─ Session Status ─────────────────────────────────────────┐
│ Project: [project name] ([project type])                 │
│ Profile: [if detected]                                   │
│                                                          │
│ SKILLS AVAILABLE                                         │
│   [icon] [skill name] ([token estimate]) - [source]      │
│   ...                                                    │
│                                                          │
│ AUTO-INVOKE RULES (from CLAUDE.md)                       │
│   [context] → [action]                                   │
│   ...                                                    │
│                                                          │
│ COMMANDS AVAILABLE                                       │
│   /[command] - [description]                             │
│   ...                                                    │
│                                                          │
│ MCP SERVERS                                              │
│   [server name] - [status]                               │
│   ...                                                    │
│                                                          │
│ ISSUES                                                   │
│   ⚠ [warning description]                               │
│   ...                                                    │
└──────────────────────────────────────────────────────────┘

How to Gather Information

  1. Project detection: Read current working directory, check for package.json, go.mod, etc.

  2. Skills: Check both locations:

    • Global: ~/.claude/skills/
    • Project: ./.claude/skills/
  3. Auto-invoke rules: Parse CLAUDE.md for tables with "Context" and "Action" columns

  4. Commands: Check both locations:

    • Global: ~/.claude/commands/
    • Project: ./.claude/commands/
  5. Token estimates: ~4 chars per token, read each skill file

  6. Issues to detect:

    • Skill referenced in CLAUDE.md but not found
    • Broken symlinks
    • Duplicate skill names (project vs global)

Example Output

code
┌─ Session Status ─────────────────────────────────────────┐
│ Project: d3-smr (D3 Visualization)                       │
│ Profile: d3-development                                  │
│                                                          │
│ SKILLS AVAILABLE                                         │
│   ⚡ d3 (1.2k) - project/.claude/skills                   │
│   ⚡ react-state (890) - project/.claude/skills           │
│   ⚡ typescript (1.1k) - project/.claude/skills           │
│   🔒 security-mindset (1.4k) - ~/.claude/skills         │
│   🔒 owasp (1.6k) - ~/.claude/skills                    │
│   📐 devlens (3.8k) - ~/.claude/skills                  │
│                                                          │
│ AUTO-INVOKE RULES                                        │
│   D3.js or visualization → /d3                           │
│   React/JSX/TSX files → /react-state                     │
│   Auth, login, passwords → /security-mindset             │
│   SQL, user input, APIs → /owasp                         │
│                                                          │
│ COMMANDS AVAILABLE                                       │
│   /session-status - This command                                 │
│   /viz/* - D3 visualization commands                     │
│                                                          │
│ MCP SERVERS                                              │
│   linear - enabled                                       │
│   sequential-thinking - enabled                          │
│                                                          │
│ ISSUES                                                   │
│   ✓ No issues detected                                  │
└──────────────────────────────────────────────────────────┘

Icons

  • ⚡ Canon/domain skill
  • 🔒 Security skill
  • 📐 Meta/methodology skill
  • 🔧 Utility skill
  • ⚠ Warning
  • ✓ OK