Wrap Up

Wrap Up

Records what was done in the current session and what needs to be done next, saving to a per-topic markdown file that accumulates across sessions.

Defaults (no config file needed):

• Output directory: wrap-up/
• File naming: wrap-up/{topic-name}.md (topic = main feature/subject worked on)
• Sections: Done, Decisions (optional), Issues (optional), Next
• Language: read from config.yaml (language field). If en, always write in English regardless of conversation language. If ko, always write in Korean. If missing, auto-detect from conversation.

---

Execution Algorithm

Step 1: Detect Topic & Match Existing File

1. Analyze conversation to identify the primary topic or feature
2. Scan existing files: Glob("wrap-up/*.md") to list all wrap-up files
3. Match logic:

Scenario	Action
Exact match found (filename = topic)	Ask user to confirm: "기존 wrap-up/{name}.md에 이어서 기록할까요?"
Similar match found (related title/content)	Show candidates and ask user to select
Multiple candidates	Present list with AskUserQuestion for selection
No match	Ask user to confirm new file creation with suggested topic name

Matching criteria:

• Filename similarity (e.g., topic auth matches auth.md)
• Title in file header (e.g., # Auth Module - Wrap Up)
• Content overlap (recent Done/Next items relate to current session's work)

Naming rules:

• Use kebab-case: business-avengers, wrap-up, api-refactor
• Be specific to the feature, NOT the project directory name

---

Step 2: Load Context (Prepend only)

If adding to an existing file:

1. Read the entire file
2. Parse the FIRST (most recent) ### Next section to identify pending items
3. Cross-reference with current session's work to determine which Next items were completed
4. This context informs Step 3 (Analyze Conversation)

---

Step 3: Analyze Conversation

Language enforcement: Read config.yaml for the language field.

• language: en → Write ALL content (section headers, bullet points, context, decisions, issues, next items) in English, regardless of conversation language.
• language: ko → Write in Korean.
• Missing or other value → Auto-detect from conversation language.

Review the entire conversation history to extract:

Section	What to Extract
Done	Tasks completed, features added, bugs fixed, refactoring done. Use conventional commit prefixes (feat, fix, refactor, docs, chore). If items from previous Next were completed, include them with "(from previous Next)" note
Decisions	Architecture choices, library selections, approach decisions made during the session
Issues	Blockers, errors, unresolved problems, workarounds applied
Next	Pending tasks, follow-up items, explicitly mentioned TODOs. Use checkbox format - [ ]

---

Step 4: Create or Append

If NEW file (user confirmed new topic):

• Create wrap-up/ directory if needed
• Write new file with topic header + session entry

If EXISTING file (user selected existing file):

• Update the FIRST (most recent) session's Next checkboxes: completed items [ ] → [x]
• Insert new session entry AFTER the header block (after > **Scope**: line), BEFORE existing sessions
• Add --- separator between the new entry and the previous most-recent session
• This ensures newest session is always at the top, oldest at the bottom (reverse chronological order)

Session date format:

• Must run date '+%Y-%m-%d %H:%M' to get the exact current time. Never estimate or guess the time.
• Format: ## Session: 2026-02-23 14:00
• Time helps identify and trace specific sessions across conversation history

Context line:

• Each session entry includes > **Context**: {brief summary} right after the session header
• 1-line summary of what the session was about (for quick identification when scanning the file)

Session ordering: Reverse chronological (newest first, oldest last).

Template (new file):

# {Topic Name} - Wrap Up

> **Project**: `{CWD}`
> **Scope**: `{relative path to primary working directory}` (e.g., `plugins/business-avengers/`)

## Session: 2026-02-23 14:00

> **Context**: OAuth 2.0 소셜 로그인 연동 및 Google/GitHub 프로바이더 구현

### Done
- ...

### Decisions
- ...

### Next
- [ ] ...

Template (existing file — insert new session at top):

# {Topic Name} - Wrap Up

> **Project**: `{CWD}`
> **Scope**: `...`

## Session: 2026-02-24 10:00        ← NEW (inserted here)

> **Context**: ...

### Done
- ...

### Next
- [ ] ...

---

## Session: 2026-02-23 14:00        ← PREVIOUS (pushed down)

> **Context**: ...

### Done
- ...

### Next
- [x] completed item (from previous Next)
- [ ] remaining item

Section omission rules:

• Omit Decisions if no significant decisions were made
• Omit Issues if no problems were encountered
• Done and Next are always included

---

Step 5: Confirm to User

Show the user:

• Full path of saved file (new or updated)
• Whether it was a new file or appended to existing
• Number of items in each section
• (Append only) Number of previous Next items completed

---

Step 6: Blog Log Generation (Optional)

After confirming the wrap-up file, check config.yaml for blog_log.enabled:

config = Read("config.yaml")  // from skill directory

if config.blog_log.enabled == false:
    exit  // skip silently

// Prompt user
AskUserQuestion(
  "블로그 로그도 생성할까요?",
  options=[
    { label: "네", description: "오늘 작업 내용을 블로그 logs 컬렉션에 저장합니다" },
    { label: "아니요", description: "wrap-up만 저장하고 종료합니다" }
  ]
)

if answer == "네":
    // Invoke wrap-to-blog skill with current session context
    // Pass: session date, topic name, Done items, Decisions, Next items
    invoke_skill("wrap-to-blog", {
      session_date: current_date,        // YYYY-MM-DD
      topic: current_topic,              // e.g., "planning-interview"
      done: session.done_items,
      decisions: session.decisions,
      next: session.next_items,
      context_summary: session.context,
      blog_dir: config.blog_log.blog_dir,
      collection: config.blog_log.collection
    })

Note: This step only runs if blog_log.enabled: true in config.yaml. If config.yaml is missing or blog_log section is absent, skip silently.

---

Trigger Phrases

English:

• "wrap up", "wrap-up", "session summary", "document progress", "record what we did"

Korean:

• "작업 정리", "세션 정리", "마무리", "진행 상황 기록", "오늘 한 일 정리"

---

Quick Reference

When to Use

Use this skill when:

• End of a work session
• Switching to a different project or topic
• Before closing a chat session
• After completing a significant milestone

Skip when:

• Very short Q&A (nothing substantial to record)
• Pure research/exploration with no actionable output

---

Error Handling

Scenario	Response
Output directory doesn't exist	Create it with mkdir -p
Write permission denied	Error: Cannot write to {path}. Check permissions.
Empty conversation	Warning: Not enough content to summarize. Continue working first.
Existing file is malformed	Append new session entry at the end regardless