Now live on npm

Catch AI scope drift before it merges.

aioengine adds AI change-control guardrails to your repo. Run it locally, in CI, or on GitHub pull requests to catch risky or out-of-scope AI-generated code before it gets committed or merged.

View on npm See commands

terminal

$npx -y aioengine@latest init --github

aioengine Init

AI change control for developers

Guardrails created

GitHub workflow

AI rule files

Snapshot guard

✓ .aioengine/config.json

✓ CLAUDE.md

✓ .cursor/rules/aioengine.mdc

✓ .github/workflows/aioengine.yml

✓ .aioengine/snapshots/.gitignore

aioengine sets up local guardrails and a pinned GitHub Actions workflow, so AI-generated changes can be checked before they merge.

AI change report

Generated before commit or on a pull request

Needs review

Files changed

Scope drift

Risky file

✓

README.md

in scope for docs update

packages/cli/package.json

possible scope drift and dependency-sensitive file

→

Recommendation

Check the flagged file before merging and keep anything unrelated out of this change.

What aioengine adds

A focused first-pass review for AI-generated changes.

It highlights scope drift, risky files, and review-heavy edits so you know exactly where to look before committing.

The local AI review toolkit

Start with one command. Add guardrails as you go.

aioengine is built for the moment after AI changes your code and before you trust it enough to commit.

Init

npx -y aioengine@latest init --github

Set up local AI coding boundaries, Claude/Cursor rules, and GitHub PR reports in one command.

Check

npx -y aioengine@latest check

Review your repo setup for missing guardrails, risky files, missing GitHub Actions, and weak review signals.

Snapshot

npx -y aioengine@latest snapshot --name checkpoint

Save Git HEAD, file hashes, and repo context before a larger AI-assisted change.

Scope

npx -y aioengine@latest scope "example: update README docs" --profile <profile>

Tell aioengine what kind of task AI was supposed to do, then flag files outside that scope. Available profiles: ui, docs, cli, ci, backend, marketing

Review

npx -y aioengine@latest review

Review current Git changes for sensitive files, dependency edits, config changes, and risky areas.

Open a PR or local CI report

npx -y aioengine@latest ci --report aioengine-report.md

Generate a Markdown report for GitHub Actions, PR comments, and review artifacts.

Why it matters

AI makes code faster. Review gets harder.

A small prompt can create a big diff. aioengine gives you a simple safety pass before you commit changes from AI coding tools.

"Snapshot" to save your current repo state

Before a larger AI edit, aioengine can save a local checkpoint with Git HEAD, tracked file hashes, and key repo context. That gives you a clear before-state, so changes are easier to inspect, compare, and roll back if an AI edit goes off track, rewrites the wrong files, or deletes large sections unexpectedly.

npx -y aioengine@latest snapshot --name checkpoint

AI can drift outside the prompt

A docs request can turn into package, workflow, auth, or config changes if you are not watching the diff closely.

Bigger diffs are harder to trust

AI tools move fast, but reviewing a wide diff across unrelated files still takes real human attention.

Risky files need extra attention

Package files, workflows, auth, billing, env files, and backend config should not slip through as casual edits.

Snapshots make reviews easier

A snapshot records Git HEAD, tracked file hashes, and repo context before AI changes code, giving you a known checkpoint.

Simple workflow

Review AI-generated code before it lands.

aioengine works best when each branch or pull request has one clear task type.

Install guardrails

Create local AI coding rules, a root .gitignore, snapshot safety files, and optional GitHub Actions.

npx -y aioengine@latest init --github

Snapshot first

Save a checkpoint before a larger AI edit so the review starts from a known repo state.

npx -y aioengine@latest snapshot --name checkpoint

Keep one task type

Use one focused branch or PR when possible. Choose a profile like ui, docs, cli, ci, backend, or marketing.

npx -y aioengine@latest scope "example: update README docs" --profile <profile>

Review the diff

Run review before committing so risky files, dependency edits, and config changes are easier to notice.

npx -y aioengine@latest review

Open a PR or save the results as a Markdown report

aioengine runs in GitHub Actions and posts the change-control report directly on the pull request. Use the command below to save the results as a Markdown report for local review or CI artifacts.

npx -y aioengine@latest ci --report aioengine-report.md

Commit with confidence

Use the report to review the files that need attention, keep unrelated changes out, and commit the focused diff.

git commit -m "Describe the focused change"

Free public CLI

Free guardrails for AI-assisted developers.

aioengine is free to start. No account required. No dashboard required. Run it locally or add it to GitHub Actions for pull request reports.

Free

Local and GitHub PR checks for AI-assisted developers.

Live now

✓Local repo setup checks
✓Root .gitignore creation
✓Claude Code guardrails
✓Cursor rules
✓Snapshot checkpoints
✓Scope profiles
✓Risky file review
✓GitHub Actions workflow
✓Pinned workflow version
✓Markdown CI reports
✓GitHub PR comments

Start on npm Send feedback

Coming soon

Future features

Paid features will focus on teams, history, policy management, and workflow visibility. The public CLI remains the starting point.

→Saved report history
→Custom project rules
→Team policy presets
→Dashboard summaries
→Multiple repo views
→Slack or email alerts
→Upgrade workflow helper
→More AI tool-specific integrations

Try the public CLI

Run aioengine in your repo now.

No account required. No dashboard required. Start with a local check, then add aioengine to GitHub Actions for pull request guardrails.

$npx -y aioengine@latest init --github

Early feedback welcome

Tried aioengine? Tell me what felt confusing.

aioengine is early and intentionally opinionated. If you hit a bug, confusing output, a false positive, or a missing feature, please open a GitHub Issue so it can be tracked publicly.

Helpful feedback includes the AI coding tool you used, the command you ran, what aioengine flagged, and whether the result felt useful or too strict.

Report bugs or request features

✓bugs or crashes
✓confusing command output
✓false positives or missed risky changes
✓feature requests

Open a GitHub Issue