# AI Agent Skipped-Step Audit Checklist

**OperatorForge AI · 2026-05-18**

A field checklist for catching AI agents that finish without completing every required step.

## 10-minute skipped-step audit

1. Restate the required outcome.
2. List every required artifact.
3. Map each requirement to evidence.
4. Check for silent substitutions.
5. Check negative constraints.
6. Confirm verification was direct.
7. Identify skipped-decision points.
8. Run a second-pass diff.
9. Record final state: done, partial, blocked, or unsafe.
10. Create one prevention rule.

## Requirement-to-evidence matrix

| Requirement | Expected evidence | Evidence found | Status | Fix needed |
|---|---|---|---|---|
| Deliverable created | File path + non-empty content |  | Pass / Fail |  |
| Public page updated | Live URL + visible change |  | Pass / Fail |  |
| Download works | HTTP 200 + correct filename |  | Pass / Fail |  |
| Brand separation preserved | Search result or grep output |  | Pass / Fail |  |
| Tests/checks run | Command + result |  | Pass / Fail |  |
| Blockers documented | Needs-from-owner note |  | Pass / Fail |  |

## Prevention prompt

```text
Before you finish, create a requirement-to-evidence table. For every requested deliverable or constraint, include the file path, command output, public URL, API read, screenshot, or exact blocker. If any requirement lacks evidence, keep working or mark it blocked with the smallest next ask. Do not claim verification from memory.
```

## Monetization angle

Offer this as an Agent Work QA Review for founders and automation consultants using AI agents in production workflows.