Pawarta bisa nyepati kutu, nglakoni tes, lan ngikat nalika sampeyan tetep ngontrol?

Question

Accepted Answer

Pawarta **agent** bisa nyopir perbaikan kutu kanthi lengkap — puter maneh, golek, ndandani, verifikasi, ikat — nanging sampeyan tetep ngontrol kanthi nyetul loop nganggo **ijin lingkup, persetujuan manungsa, lan langkah-langkah cilik sing bisa dibalekake**. Agent ngetik; sampeyan mutusake apa sing sejatine teka.

## Loop

```text
1. REPRODUCE  → write/run a failing test that demonstrates the bug (red)
2. LOCATE     → search the codebase, read the relevant files, form a hypothesis
3. PROPOSE    → draft a minimal fix (you can review the plan before any edit)
4. VERIFY     → run tests + lint/typecheck → fix is proven by the test going green
5. REVIEW     → show the DIFF for your approval (nothing committed yet)
6. COMMIT     → only after approval, on a branch, with a clear message
```

Gawe penting yaiku **verifikasi didorong tes**: agent ora ngaku "dandi" — dheweke nedha kanthi nggaweke tes sing gagal dadi lulus. Tes sing gagal saka langkah 1 yaiku sinyal objektif manawa owahan wis maido.

## Mekanisme kontrol

```text
- SCOPED PERMISSIONS → allowlist safe commands (test, lint); prompt for the rest
- APPROVAL GATES     → human confirms destructive/irreversible actions + the final commit
- SMALL STEPS        → one focused change at a time → easy to review and revert
- BRANCH / WORKTREE   → work isolated from main; throw it away if it goes wrong
- DIFF REVIEW        → read the actual diff before commit, not just the agent's summary
```

```bash
# The agent runs the suite itself and reports the result before proposing a commit
npm test -- --runInBand   # red before the fix, green after → objective verification
git diff                  # you review every line that will be committed
```

Cawe-cawe ing **cabang utawa git worktree** tegese kasus paling elek yaiku `git checkout main` — eksperimen ora nernah ngiloni kaanan kerjane sampeyan. Commit cilik sing bisa dibalekake nggaweke langkah siji gampang diperiksa lan dibatalake.

## Mangapa penting

Cawe agentic nggeser sampeyan saka ngetik saben baris dadi **ninjau lan ngarahake**. Daya ungkile nyata, nanging risiko agent kanthi yakin ngikat owahan salah utawa musnahake uga nyata. Tetep mendhak prentah persetujuan manungsa sadurunge ikat, verifikasi kanthi tes tinimbang percaya prosa, lan pisahhake gawe ing cabang nyinaukan sampeyan nangkep kecepetan otomasi nalika tetep sampeyan sing mutusake apa sing teka ing `main`.