AgentSkillsCN

session-wrap

会话总结与多代理分析。当用户说“结束”、“wrap”、“会话结束”、“结束会话”、“我应该提交什么”或希望在结束编码会话前分析已完成的工作时使用。建议更新文档、寻找自动化机会并安排后续任务。

SKILL.md
--- frontmatter
name: session-wrap
description: "Session wrap-up with multi-agent analysis. Use when the user says 'wrap', '/wrap', 'session wrap', 'end session', 'what should I commit', or wants to analyze completed work before ending a coding session. Suggests documentation updates, automation opportunities, and follow-up tasks."

Session Wrap-up

Comprehensive session wrap-up workflow with multi-agent analysis.

Quick Usage

  • /wrap - Interactive session wrap-up (full workflow)
  • /wrap [message] - Quick commit with provided message

Execution Flow

code
1. Check Git Status
2. Phase 1: 4 Analysis Agents (Parallel)
   ┌─────────────┬──────────────┐
   │ doc-updater  │ automation-  │
   │              │ scout        │
   ├─────────────┼──────────────┤
   │ learning-    │ followup-    │
   │ extractor    │ suggester    │
   └─────────────┴──────────────┘
3. Phase 2: duplicate-checker (Sequential)
4. Integrate Results & AskUserQuestion
5. Execute Selected Actions

Step 1: Check Git Status

bash
git status --short
git diff --stat HEAD~3 2>/dev/null || git diff --stat

Step 2: Phase 1 - Analysis Agents (Parallel)

Prepare a session summary, then execute 4 agents in parallel (single message with 4 Task calls).

Session Summary (provide to all agents)

code
Session Summary:
- Work: [Main tasks performed]
- Files: [Created/modified files]
- Decisions: [Key decisions made]

Agent Roles

AgentRoleOutput
doc-updaterAnalyze CLAUDE.md/context.md updatesSpecific content to add
automation-scoutDetect automation patternsSkill/command/agent suggestions
learning-extractorExtract learning pointsTIL format summary
followup-suggesterSuggest follow-up tasksPrioritized task list

Step 3: Phase 2 - Validation Agent (Sequential)

Run after Phase 1 completes. Feed all Phase 1 results to duplicate-checker:

  1. Complete duplicate → Recommend skip
  2. Partial duplicate → Suggest merge approach
  3. No duplicate → Approve for addition

Step 4: Integrate Results

Present combined analysis:

markdown
## Wrap Analysis Results

### Documentation Updates
[doc-updater summary + duplicate-checker feedback]

### Automation Suggestions
[automation-scout summary + duplicate-checker feedback]

### Learning Points
[learning-extractor summary]

### Follow-up Tasks
[followup-suggester summary]

Step 5: Action Selection

code
AskUserQuestion(
    question: "Which actions would you like to perform?",
    multiSelect: true,
    options: [
        "Create commit (Recommended)",
        "Update CLAUDE.md",
        "Create automation",
        "Skip"
    ]
)

Step 6: Execute Selected Actions

Execute only the actions selected by user.

When to Skip

  • Very short session with trivial changes
  • Only reading/exploring code
  • Quick one-off question answered

Additional Resources

  • See references/multi-agent-patterns.md for detailed orchestration patterns
  • Agent definitions in agents/ directory