When should the skill be used?

After making backend (Python) or frontend (TypeScript) changes, before committing or finishing a work session.

Does it work on committed files?

No. It runs its checks on all uncommitted files.

Checking Changes

SkillDev tools

Checking-changes is a skill that validates code changes before committing. When invoked, it has the agent run 'make check', which formats, lints, type-checks, scans for unused exports and dependencies, and runs unit tests on all uncommitted files. It repeats the process, fixing reported errors each time, until every check passes.

Use Checking Changes in Claude, ChatGPT or Ahel Desktop

Free. Sign in, add Checking Changes and connect your AI. About a minute.

Claude ChatGPT Ahel Desktop

Also: Claude Code · Cursor · Codex

Then ask your AI: use the Checking Changes skill

Details

Instructions available. Your AI can read the instructions. Execution depends on the setup they require.

Account requirements not reviewed. Check the skill instructions before use; ahel provides instructions and does not run this skill.

No other account needed.

Have a project with a 'make check' target that runs format, lint, type, and unit test checks.

Checking ChangesStart free

What your AI can do with it

Runs 'make check' on all uncommitted files
Formats and lints Python and TypeScript changes
Type-checks backend (Python) and frontend (TypeScript) code
Scans for unused exports and dependencies
Runs unit tests on uncommitted files
Repeats checks and fixes reported errors until all pass

Getting started

Have a project with a 'make check' target that runs format, lint, type, and unit test checks.
Add the checking-changes skill to your agent setup.
Make your backend (Python) or frontend (TypeScript) changes.
Ask the agent to use the skill before committing or finishing a work session.

What this skill tells your AI

The instructions your AI receives, as published by streamlit/streamlit in .claude/skills/checking-changes/SKILL.md and read by ahel’s review.

Run at the end of a work session or after completing a set of changes — not after every small edit.

make check

This runs formatting, linting, type checking, Knip unused-export and unused-dependency analysis, and unit tests on all uncommitted files (staged, unstaged, and untracked).

Workflow

Run make check
If issues are found:
- Fix the reported errors
- Re-run make check
- Repeat until all checks pass
Only consider work complete when make check succeeds

Notes

E2E tests are not included by default; use E2E_CHECK=true make check to also run changed e2e tests
E2E snapshot mismatches can be ignored (they require manual updates)

Signals

GitHub stars: 46k
Forks: 4k
Last commit: Oct 2026

Questions

When should the skill be used?: After making backend (Python) or frontend (TypeScript) changes, before committing or finishing a work session.
What does the skill actually run?: It has the agent run 'make check', which formats, lints, type-checks, scans for unused exports and dependencies, and runs unit tests on all uncommitted files.
What happens if a check fails?: The skill repeats the process, fixing reported errors each time, until every check passes.
Does it work on committed files?: No. It runs its checks on all uncommitted files.

Advanced

Item type: skill
Key: checking-changes
Source: github.com/streamlit/streamlit

github.com/streamlit/streamlit