@codewithmukesh/checkpoint

@codewithmukesh/checkpoint — AI coding skill

View in AI SkillSafe app
0 downloads
0 stars
0 demos
SKILL.md
namecheckpoint
description>
disable-model-invocationtrue

/checkpoint

What

A quick mid-session save that banks the known-good state in two moves:

  1. Descriptive git commit — stage relevant changes and commit with a message that summarizes the work.
  2. Brief handoff note — write .claude/handoff.md so a resumed session (or you, after a botched refactor) knows exactly where things stand.

Checkpoint is the mid-session save; /wrap-up is the end-of-session ritual. Checkpoint commits and jots a note, then keeps working. Wrap-up does the full handoff plus learning extraction into MEMORY.md.

When

  • Before a risky refactor or destructive change — checkpoint the known-good state
  • When switching to a different task or feature mid-session
  • After completing a logical unit of work — bank it
  • User says "checkpoint", "save progress", "save state", "pause here"
  • If the session is actually ending, use /wrap-up for the full ritual

How

  1. Assess stategit status and git diff to see staged and unstaged changes.
  2. Stage changes — Stage specific files; never git add -A. Exclude generated files, build artifacts, and secrets (.env, credentials).
  3. Commit — Descriptive message following the repo's commit style (conventional prefixes: feat:, fix:, refactor:). New commit, never amend. If there is nothing to commit, skip to the handoff.
  4. Write handoff note.claude/handoff.md, same format as /wrap-up but abbreviated: Completed, Pending (with specific next steps), Learned (only if something non-obvious surfaced), and Context (branch + commit hash). One file, always overwritten.
  5. Confirm — Report the commit hash and the handoff summary, then resume work.

Safety Rules

  • Never push, force-push, or touch remote branches
  • Never commit secrets
  • Never amend — a checkpoint is a new commit you can return to
  • No changes to commit? Update the handoff only

Example

User: /checkpoint — about to try the EF Core split-query refactor

Claude: 3 files modified, 1 new.

  Commit a1b2c3d: "feat: add CreateOrder validation with FluentValidation"
  - src/Orders/Features/CreateOrder.cs
  - src/Orders/Features/CreateOrderValidator.cs (new)
  - tests/Orders.Tests/Features/CreateOrderTests.cs

  Handoff updated (.claude/handoff.md):
  ## Completed
  - [x] CreateOrder validation + 3 test scenarios
  ## Pending
  - [ ] Split-query refactor in GetOrderDetails (about to attempt)
  ## Context
  - Branch: feature/order-validation | Checkpoint: a1b2c3d

  Checkpoint saved — safe to proceed with the risky refactor.

Related

  • /wrap-up — End-of-session ritual: full handoff format definition plus learning extraction into MEMORY.md
  • /build-fix — Get the build green before checkpointing

Embed badges

Add these to your README to show the skill's verification status.

SkillSafe verified badge
Verified badge
[![SkillSafe verified badge](https://api.skillsafe.ai/v1/badge/@codewithmukesh/checkpoint/verified)](https://skillsafe.ai/skill/@codewithmukesh/checkpoint/)
Installs badge
Installs badge
[![Installs badge](https://api.skillsafe.ai/v1/badge/@codewithmukesh/checkpoint/installs)](https://skillsafe.ai/skill/@codewithmukesh/checkpoint/)
Scan badge
Scan badge
[![Scan badge](https://api.skillsafe.ai/v1/badge/@codewithmukesh/checkpoint/scan)](https://skillsafe.ai/skill/@codewithmukesh/checkpoint/)
Eval pass rate badge
Eval pass rate
[![Eval pass rate badge](https://api.skillsafe.ai/v1/badge/@codewithmukesh/checkpoint/eval)](https://skillsafe.ai/skill/@codewithmukesh/checkpoint/)