Your AI agent uses your API keys without the keys landing in its context.

Secrets stay in macOS Keychain. Claude Code gets back ok or failed, nothing else. EnvSecure is a Claude Code mod from Syncoders Labs.

A Claude Code plugin for macOS.

Get EnvSecure

What the AI sees, and what it does not

The agent sees

  • The name of each secret and your description of it.
  • envsecure: ok or envsecure: failed after a run.
  • Output with the value redacted, only if you switch "AI can see" on for that secret. It resets every session.

The agent does not see

  • The secret value, in normal use.
  • The command output, the exact exit code or its size.
  • The full run log. You read it in the pane.

What it does not do

EnvSecure stops accidental leaks into chat, transcripts and logs. It is not a cage for a hostile model, and we say so.

Install

After you buy, your receipt email carries a zip file. Unzip it, then run two commands in a terminal, and finish in the pane.

claude plugin marketplace add <folder>
claude plugin install envsecure@syncoders-lab

Replace <folder> with the unzipped folder. Then paste your license key in the EnvSecure pane. macOS only. Claude Code only.

Pricing

One key works on up to 5 of your Macs. Cancel anytime.

Monthly

$4 per month
Buy monthly

Yearly

$36 per year

3 months free.

Buy yearly

License key by email, 5 Macs per key.

The licence runs while your subscription is paid. You accept the License and Terms at checkout.

Questions

Can I cancel?

Cancel anytime.