tejasgadhia

tg-recap

0
0
# Install this skill:
npx skills add tejasgadhia/tg-claude-skills --skill "tg-recap"

Install specific skill from multi-skill repository

# Description

Use at end of coding sessions to generate comprehensive session documentation (PROJECT-STATUS.md, NEXT-STEPS.md, LEARNINGS.md) with git commit suggestions and next-session prompts. Wrap up sessions for continuity.

# SKILL.md


name: tg-recap
description: Use at end of coding sessions to generate comprehensive session documentation (PROJECT-STATUS.md, NEXT-STEPS.md, LEARNINGS.md) with git commit suggestions and next-session prompts. Wrap up sessions for continuity.


Session Recap - End of Session Documentation

Overview

Generate comprehensive session documentation to bridge stateless AI sessions and preserve context for future work. Designed for Tejas's workflow: capture decisions, learnings, and next steps clearly.

When to Use

Always use at the end of a coding session when:
- You've completed significant work
- You're wrapping up for the day
- User says "that's it for now" or similar
- Before switching to a different project

Trigger phrases:
- "/recap"
- "Wrap up this session"
- "Document what we did"
- "Prepare for next session"
- "That's all for today"

The Workflow

Step 1: Analyze Session Context

Gather information about what happened:

# Check repository status
git status

# See what changed
git diff --stat

# Review recent commits
git log --oneline -10

# Check uncommitted changes
git diff

Identify:
- New files created
- Files modified
- Files deleted
- Uncommitted changes
- Key decisions made during session
- Technical patterns discovered
- Problems solved
- Rationale behind major choices

Extract from conversation history:
- What did we build?
- What problems did we solve?
- What decisions did we make (and why)?
- What approaches did we try?
- What worked well?
- What didn't work?

Step 2: Generate PROJECT-STATUS.md

Purpose: Snapshot of project state for next session

# Project Status - [Project Name]
Last Updated: [Date]

## Current State

**What's Working:**
- [Feature/component that's complete and working]
- [Another completed item]

**In Progress:**
- [What's partially done]
- [What needs finishing]

**Not Started:**
- [Planned features not yet begun]

## Recent Changes

### [Date] Session
- [What was built/fixed]
- [Technical decisions made]
- [Files created/modified]

## Architecture

**Tech Stack:**
- [List current technologies]

**Structure:**
- [Key files and their purposes]
- [How components relate]

## Known Issues

- [Bug or limitation]
- [Another issue]

## Next Priorities

1. [Most important next task]
2. [Second priority]
3. [Third priority]

Step 3: Generate NEXT-STEPS.md

Purpose: Clear action items for next session

# Next Steps - [Project Name]
Last Updated: [Date]

## Immediate Tasks (Start Here)

### 1. [Task Name]
**Priority**: HIGH
**File(s)**: `file.ext` lines X-Y
**What to do**: [Specific steps]
**Why**: [Context/reasoning]
**Estimated effort**: [Quick/Medium/Substantial]

### 2. [Another Task]
**Priority**: HIGH
[Same structure]

## Future Enhancements

- [Nice-to-have feature]
- [Polish item]
- [Optimization]

## Questions to Resolve

- [Unanswered question]
- [Decision needed]

## Blockers

- [What's blocking progress, if anything]

## Next Session Starter Prompt

Copy this to start your next session:

> "Continue working on [project-name]. Last session: [brief summary of what was done]. Next: [top priority from immediate tasks]. Reference NEXT-STEPS.md for details."

Step 4: Generate LEARNINGS.md

Purpose: Capture insights and patterns to reuse

# Learnings - [Project Name]
Last Updated: [Date]

## What Worked Well

### [Pattern/Approach Name]
**Context**: [When this happened]
**What we did**: [Specific approach]
**Why it worked**: [Reasoning]
**Reuse for**: [Future scenarios where this applies]

## What Didn't Work

### [Attempted Approach]
**What we tried**: [Specific attempt]
**Why it failed**: [Root cause]
**What we did instead**: [Solution]
**Lesson**: [What to remember for next time]

## Technical Patterns

### [Pattern Name]
**Implementation**: [Code snippet or approach description]
**Use case**: [When to use this]
**Benefits**: [Why this is good]

## Efficiency Improvements

### Better Questions to Ask Upfront
- [Question that would have saved iteration]
- [Another clarifying question]

### Workflow Optimizations
- [Process improvement discovered]
- [Tool/technique that sped things up]

## Dependencies & Tools

### [Library/Tool Name]
**What it does**: [Purpose]
**Why we chose it**: [Decision rationale]
**Gotchas**: [Things to watch out for]
**Alternatives considered**: [What we didn't use and why]

## Retrospective

**What went well:**
- [Success point]

**What could have been faster:**
- [Efficiency opportunity]

**What to do differently next time:**
- [Process improvement]

Step 5: Suggest Git Commits

If there are uncommitted changes:

## Suggested Git Commits

Based on changes made this session:

### Commit 1: [Brief description]
**Files**: [List changed files]
**Message**:

[Commit message following project conventions]

Changes:
- [What this commit includes]
- [Another change in this logical group]

Co-Authored-By: Claude Sonnet 4.5 [email protected]

**Command**:
```bash
git add [files]
git commit -m "[message]"

[Repeat for logical commit groups]

### Step 6: Output Summary

**Present to user:**

```markdown
# Session Recap Complete

## Documents Created/Updated

✅ **PROJECT-STATUS.md** - Current project state snapshot
✅ **NEXT-STEPS.md** - Action items for next session
✅ **LEARNINGS.md** - Insights and patterns captured

## Uncommitted Changes

[If changes exist]:
- [N files] modified
- [N files] created
- Suggested [N commits] to organize changes

[If no changes]:
✅ All changes committed

## Next Session

**Start with this prompt:**
> [Next session starter prompt from NEXT-STEPS.md]

**Priority**: [Top item from NEXT-STEPS.md]

---

Session documentation saved. You're ready to pick up where you left off!

Key Principles

Comprehensive Context:
- Capture decisions and rationale (not just what, but why)
- Document patterns to reuse across projects
- Clear next steps so you don't have to remember

Concise but Complete:
- Bullet points over paragraphs
- Specific file references (file.ext:line)
- Actionable items, not vague goals

Continuity-Focused:
- Next-session prompt should have all context needed
- Prioritize tasks clearly (what to do first)
- Highlight blockers upfront

Learning-Oriented:
- Document what worked AND what didn't
- Capture efficiency improvements
- Questions to ask earlier next time

Success Criteria

PROJECT-STATUS.md created - Clear snapshot of project state
NEXT-STEPS.md created - Prioritized action items with context
LEARNINGS.md created - Patterns and insights documented
Git commits suggested - Logical groupings if uncommitted changes
Next-session prompt - Ready to copy/paste for continuity
Efficiency insights - Questions/workflows to improve captured

Example Output

# Session Recap Complete

## Documents Created/Updated

✅ **PROJECT-STATUS.md** - Flight tracker MVP complete, charts working
✅ **NEXT-STEPS.md** - 3 HIGH priority fixes from review + 5 enhancements
✅ **LEARNINGS.md** - Chart.js theming pattern, CSV parsing approach

## Uncommitted Changes

- 3 files modified (index.html, styles.css, app.js)
- Suggested 1 commit: "Add MVP flight list and Chart.js visualizations"

## Next Session

**Start with this prompt:**
> "Continue working on flight-tracker. Last session completed MVP (flight list + charts). Next: Fix HIGH priority issues from review - dark mode contrast (styles.css:45-60), show filters upfront (index.html:89), CSV error handling (app.js:145). Reference NEXT-STEPS.md for details."

**Priority**: Fix dark mode contrast failures (WCAG AA compliance)

---

Session documentation saved to ~/Claude/flight-tracker/

Ready for your next session!

Real-World Impact

Before this skill: Context lost between sessions, forgot decisions, repeated questions, unclear what to do next.

After this skill: Full context preserved, decisions documented, clear priorities, pick up instantly where you left off.

Time saved per session transition: 15-30 minutes of context rebuilding.
Efficiency gain: No repeated questions, no forgotten approaches.

# Supported AI Coding Agents

This skill is compatible with the SKILL.md standard and works with all major AI coding agents:

Learn more about the SKILL.md standard and how to use these skills with your preferred AI coding agent.