session-handoff
Generate a complete handoff prompt for seamless continuation in a new session. Use when context is filling up or before ending a productive session.
用 Codex 或 Claude 帮你安装 复制这段 Prompt,粘贴到 Codex、Claude 或其他助手里,让它检查 Skill 页面并帮你完成安装。
菜单
Generate a complete handoff prompt for seamless continuation in a new session. Use when context is filling up or before ending a productive session.
用 Codex 或 Claude 帮你安装 复制这段 Prompt,粘贴到 Codex、Claude 或其他助手里,让它检查 Skill 页面并帮你完成安装。
基于 SOC 职业分类
| name | session-handoff |
| description | Generate a complete handoff prompt for seamless continuation in a new session. Use when context is filling up or before ending a productive session. |
Generate a complete handoff package for seamless session continuation.
See .ai/memory/README.md for how memory files fit together and default read order for the next session.
Read .ai/memory/.memory-config.md for token budget (if exists):
economy: Session summary ~200-400 tokens (brief, 2-3 sentences)light: Session summary ~400-700 tokens (concise, 3-5 sentences)standard: Session summary ~600-1000 tokens (default, 5-8 sentences)detailed: Session summary ~900-1500 tokens (comprehensive, 8-12 sentences)If no config exists, use standard budget.
Before generating handoff:
Session file format: sessions/YYYY-MM-DD-topic.md (see .ai/memory/sessions/README.md)
Note: Include timestamp (HHMM) to avoid overwriting previous same-day sessions. Use topic suffix when a clear topic exists.
# Session: [date]
## Summary
Brief summary of what was accomplished.
- economy: 2-3 sentences
- light: 3-5 sentences
- standard: 5-8 sentences (default)
- detailed: 8-12 sentences with comprehensive context
## Work Completed
- [Item 1]
- [Item 2]
## Decisions Made
- ADR-NNN: [title] (if applicable)
## Context for Next Session
- [Key context point]
## Open Questions
- [Question if any]
---
*Session duration: ~[time]*
Collect:
Output a copyable prompt:
## Handoff Prompt (copy everything below)
```
I'm continuing work on [project-name] from a previous session.
## Original Goal
[What the user originally asked for]
## Session Summary
[2-3 sentence summary]
## Current State
- **Active task:** [Current work]
- **Files in progress:** [List files]
- **Last action:** [Recent action]
## Completed This Session
- [Item 1]
- [Item 2]
## Remaining Work
- [ ] [Task 1 - next priority]
- [ ] [Task 2]
## Key Decisions Made
- [Decision 1] (see ADR-NNN)
## Open Questions/Blockers
- [Question or blocker]
## Context to Load
Project memory is at: .ai/memory/
Key files to review: [list files]
Please load the project memory and continue with [next task].
```
Session handoff complete. Memory synced.
Copy the prompt above into a new session to continue.
Work with Storybook in apps/ui (colocated stories, MSW, TanStack Start, Storybook MCP). Use when writing or updating stories, using the ctxpipe-storybook MCP, previewing components in Storybook, or when adding new UI components or pages in apps/ui.
React UI patterns for apps/ui—Effects vs rendering, TanStack Query for all server data, useMemo, keys, event handlers. Start here when creating or editing React components.
Initialize the ConKeeper memory system for the current project. Creates the directory structure and starter files for persistent AI context.
Track recurring session friction and success patterns across memory (sessions/, patterns). Use periodically or when improving agent instructions—not every turn.
Lightweight session retrospective: summarize what worked, what confused the agent, and propose a sessions/ note or pattern bullet. Use after substantial tasks or before handoff.
Search .ai/memory for keywords, topics, or ADR numbers before loading large files. Use when you need relevant memory without reading patterns.md or all ADRs whole.