메뉴
Use when user says "let's wrap up", "close shop", "done for today", or wants to end a session. Handles session wrap-up including git operations, documentation updates, roadmap updates, and preparing for next session. Works across all repo types.
Codex 또는 Claude로 설치 이 Prompt를 복사해 Codex, Claude 또는 다른 어시스턴트에 붙여 넣으면 Skill 페이지를 검토하고 설치를 진행할 수 있습니다.
SOC 직업 분류 기준
| name | wrap-up-session |
| description | Use when user says "let's wrap up", "close shop", "done for today", or wants to end a session. Handles session wrap-up including git operations, documentation updates, roadmap updates, and preparing for next session. Works across all repo types. |
Wrap up a work session. Execute these steps directly (do NOT spawn a subagent — you have the session context, a subagent does not).
Check CLAUDE.md for ## Session Wrap-up, ## Deployment, or similar sections. If the project has specific wrap-up steps (sync commands, restart commands, deploy scripts), note them — you'll run them in Step 4.
git status --short
git diff --stat
git log --oneline -5
If there are uncommitted changes, ask the user what to do:
git stash push -m "WIP: [description]" — for work-in-progress not ready to commit. /start-session will surface this next time.Push: After committing, push to the remote branch. Ask the user if unsure which branch.
Update roadmap sections — skip if CLAUDE.md lacks these sections:
## Session Log — Keep only the 1 most recent entry in CLAUDE.md. Replace (don't append) the existing session log content with today's entry:
### YYYY-MM-DD
- Completed: [infer from git commits + conversation context]
- Next: [ask user if unclear]
Move any older entries to docs/session-archive.md. Create the archive file if it doesn't exist. claude-mem already stores full session history — the session log exists only so /start-session can show "last time + next steps."
## Roadmap — Check off completed items (- [ ] → - [x]). Infer completions from commits made during the session.
## Current Focus — Update to the next incomplete roadmap item, or ask the user.
Commit and push the CLAUDE.md update.
Before writing anything to auto memory (~/.claude/projects/*/memory/), check:
If you spot existing auto memory files that duplicate CLAUDE.md content, note them in the session summary so the user can clean up (or run node ~/.claude/scripts/memory-audit.js).
Ask the user a single question: "Any surprising or non-obvious learnings from this session worth lifting into memory?"
/promote-memory — that skill mines claude-mem for recurring patterns, "searched twice" items, and cross-project themes. This step only handles fresh learnings from the current session.Keep this lightweight — don't query claude-mem here. Most sessions will answer "no" and that's fine.
If the repo uses git worktrees:
git worktree list 2>/dev/null
If other worktrees have uncommitted changes, note them in the summary so the user is aware.
Flag any background processes started during this session (nohup jobs, pm2 processes, screen sessions, drip campaigns, dev servers). Note any that are still running so the user knows what's active after the session ends.
Run whatever sync or deploy commands are documented in the project's CLAUDE.md (found in Step 0). Common examples: rsync to VPS, deploy scripts, pm2 restart. If nothing is documented, skip this step.
Provide a brief summary: what was done, what's next, any background processes still running, any uncommitted carry-forward files.
Generate a concise session title (3-8 words) that captures the main accomplishment. Format: [verb] [what] (e.g., "Deploy Copilot Studio channel", "Fix timezone bug in cron", "Build document extraction pipeline").
IMPORTANT: /rename is a built-in CLI command that only the user can execute — you cannot invoke it. Output the suggested title and ask the user to run it:
Suggested session name: **<title>**
Run: `/rename <title>`
Check CLAUDE.md for type: declaration or infer from contents.
/start-session which reads these same roadmap sectionsConvene a 3-model council (Claude + GPT via codex CLI + Gemini CLI) on a high-stakes decision. Forces cross-critique between members and surfaces where they actually disagree, breaking Claude's default agreeableness. Use when the user asks to "convene a council", "get a second opinion", "ask GPT and Gemini", "what would other models say", or has an architecture / strategy / hiring / pricing decision where being wrong is expensive. Skip for factual questions, code with one right answer, or anything premortem-shaped (route to premortem skill instead).
Architect, execute, and iterate on AI evaluations using the Data-Task-Score framework. Treats evals as the modern, quantifiable version of a PRD. Use when the user asks to "build an eval," "improve model quality," "test an agent workflow," "quantify product intuition," "move beyond vibe checks," "measure AI output," "score LLM responses," "benchmark a prompt," or "set up evaluation infrastructure." Also triggers on phrases like "how do I know if this is working," "is the model getting better," or "eval-driven development."
Design and install validation hooks for coding agents (e.g., Claude Code) to make AI changes safer and more deterministic. Use when you want post-tool-use or stop hooks, automated tests/linters/formatters, parallel subagents with per-file validation, or a repeatable “agent pipeline” with audit logs.
Apply the codebase singularity approach: reliable codebase understanding and change with a repeatable workflow, guardrails, and verification gates. Use for repo work (feature, bugfix, refactor, migration) when you need high trust, minimal diffs, and explicit validation and exit criteria.
Use when user says "let's get started", "where are we", or at beginning of a session. Reads project context from CLAUDE.md, checks git status and recent commits, and provides orientation for the session. Works across all repo types (code, research, mixed).
Extract tweet series from X/Twitter profiles with full content, links, and engagement metrics. Use when: (1) User provides a tweet URL and wants similar tweets by the same author, (2) User asks to "extract tweets", "collect tweet series", or "scrape tweets", (3) User wants to analyze a user's recurring tweet format (e.g., daily updates, weekly roundups), (4) User needs tweet content with embedded links preserved for analysis. Requires Chrome browser automation tools (mcp__claude-in-chrome__*).