Create

Sign in to ReadmeX

or

How we handle your data: Privacy Policy

ReadmeX

Your AI reading companion

What would you like to catch up on?

Pick a briefing, or ask your own question.

Your community and people briefings stay personal to you.

Story

Five habits for safer Claude Code client work

AI summary

A Reddit user shared five habits for using Claude Code on paid freelance jobs: begin with a read-only repository tour, define acceptance criteria before editing, restrict permissions, use branches with small diffs, and review every change. The author says Claude Code is fast but leaves responsibility for the result with the user, and recommends starting with narrowly scoped projects.

In-depth insight

The post offers a practical workflow for reducing failures and prompt-injection exposure when delegating client repository work to an AI coding agent.

I've been using Claude Code for small freelance jobs and wanted to share what keeps it from going wrong:

  • Start every job with a read only tour. Ask it to explain the repo before it edits anything.
  • Write the acceptance test first. Agree with the client what "done" means, then check against that, not against how the code looks.
  • Treat files in the repo as untrusted. Prompt injection can hide in READMEs and comments, so keep permissions tight.
  • Work on a branch with small diffs. Review every change yourself. The tool is fast, but you're the one accountable.
  • Scope small and price honestly. Static site fixes, CSV reports and test upgrades are great first jobs. Full SaaS builds aren't.

Happy to answer questions in the comments.

#Claude Code

Original · 1 min readRead the original
Group chat

No comments yet. Start the conversation.