Summarize Work

Generate meaningful summaries of work completed in a session. Use for extended commit messages or branch-level documentation.

When to Use

•Before committing to generate extended commit messages
•When reviewing work on a feature branch before creating a PR
•To document challenges and iterations for team context

Workflow

code

Summary Progress:
- [ ] Phase 1: Gather context (session history + git state)
- [ ] Phase 2: Categorize changes
- [ ] Phase 3: Generate summary

Phase 1: Gather Context

Analyze the session history to understand:

•What changed — Files modified, functions added/removed, APIs changed
•Why it changed — User requests, bug reports, review feedback
•How it evolved — Iterations, failed approaches, pivots

Check git state to correlate with session work:

bash

# Staged and unstaged changes
git diff --stat
git diff --cached --stat

# If on a branch, changes since divergence from main
git log --oneline origin/main..HEAD 2>/dev/null || git log --oneline origin/master..HEAD

Phase 2: Categorize Changes

Group work into these categories (skip empty categories):

Category	Description	Examples
Features	New functionality	New endpoints, UI components, CLI commands
Fixes	Bug corrections	Edge case handling, validation fixes, error corrections
Refactors	Code improvements	Extractions, simplifications, performance
Challenges	Multi-iteration problems	Issues requiring 3+ attempts, debugging sessions
Review fixes	Code review feedback	Changes made in response to review comments

Phase 3: Generate Summary

Produce a summary following this structure:

For commit messages (extended body):

code

<short description>

<One-sentence context — the problem or motivation>

Changes:
- <Major change 1>
- <Major change 2>

<Optional: Challenges section if significant iteration occurred>

For branch summaries (PR descriptions):

code

<short description>

<2-3 sentences describing the overall goal and outcome>

- <Grouped by feature/component>
- <Major change 2>

<Optional: Document any significant debugging or iteration>

<Optional: Changes made from code review feedback>

Key Guidelines

•Focus on meaningful changes — Skip trivial edits, typo fixes, or formatting
•Capture the "why" — Not just what changed but why it mattered
•Highlight challenges — Multi-iteration problems are valuable context for future maintainers
•Be specific — "Fix auth token expiry handling" not "Fix auth bug"
•Use imperative mood — "Add", "Fix", "Refactor" not "Added", "Fixed"

Output Calibration

Context	Detail Level
Single commit	1-3 bullet points, focus on the "what"
Multi-commit session	Group by theme, include challenges
Feature branch	Full summary with context and challenges

Anti-patterns

•Listing every file changed (the diff shows that)
•Including trivial changes alongside meaningful ones
•Vague descriptions ("various improvements", "code cleanup")
•Missing context for complex changes