Safety•Start Here — Build Safely With AI•10 min lesson

Safe Agent Loop

Use this before implementation work when you want the agent to set one goal, explain the plan, verify the result, and stop after one reviewable change.

Read the matching lesson

vibe-codingworkflowbeginnerssafety

Copy-ready prompt

"I want to work in a safe beginner loop.
Please do only this one task: [describe one tiny change].

Goal:
- outcome: [what should be true]
- scope: [files, pages, or systems in scope]
- proof: [test command, browser check, or manual check]
- stop condition: [when to stop]
- do not touch: [auth, payments, database, deployment, packages, secrets, etc.]

Before making changes:
1. explain your plan in plain English
2. list the files, commands, and tools you expect to use
3. tell me if this needs a stronger model, deeper planning, a subagent, MCP, or a loop -- and why
4. do not add packages or change config unless absolutely necessary

After making changes:
5. tell me exactly what changed
6. tell me how to test it in one minute
7. run only the verification command or browser check we agreed on
8. stop so I can review before the next step"

Prompt FAQ

Questions to answer before you paste it

When should I use the Safe Agent Loop prompt?

Use this before implementation work when you want the agent to set one goal, explain the plan, verify the result, and stop after one reviewable change. Use it when you need a safer starting point than a blank prompt and you want the agent to stay inside explicit constraints.

Should I paste this prompt exactly as written?

No. Treat it as a safe starter. Replace the task, files, constraints, and verification details with your actual context before you run it.

What should I do after the agent answers?

Read the diff, run the checks, and stop after one reviewable step. If you need deeper context, open the lesson that explains the reasoning behind the prompt.

Related prompts worth copying next

Start Here — Build Safely With AI

Don't Lose Your Work — Folders, Git, and Checkpoints

The minimum safe setup for total beginners: a real project folder, a Git repo, a remote backup, and repeatable checkpoints.

Open prompt page

Secrets Hygiene

Pre-Flight Secrets Check

Run this before you paste configs, screenshots, or terminal output into an AI tool so you do not leak API keys, connection strings, or internal URLs.

Open prompt page

Start Here — Build Safely With AI

What You're Actually Doing When You Build With AI Agents

A plain-English explanation of the job: modern AI agents can plan, edit, run tools, and loop, but you still choose scope, inspect output, and own the result.

Open prompt page