SafetyAppears in · Band / Toast / PaneAdded

LeakStop

Stops secrets before Claude writes, prints or commits them, and can steer it to env vars.

LeakStop · Effect preview · IllustrationIllustrative
> Add pagination to the order list
⏺ I’ll look at the list component first.
⚠ LeakStop · about to write a likely Anthropic API key
sk-ant-api03-****…x7f
[1] Block [2] Allow once [3] Use an env var
❯ keep editing the login route
❯

Illustration. Actual results depend on the mod, terminal and environment.

Install

/plugin marketplace add AlexandreMartinezOlmos/claude-code-leakstop
/plugin install leakstop@leakstop
/reload-plugins

Run these commands one at a time in a Claude Code session.

What it accesses

Based on the author’s documentation, checked when listed; use claude plugin validate to check for yourself.

Reads

Yes

Reads files about to be written or opened, .leakstop.json, and large Bash output copies under tool-results.

Writes files

Yes

Stores permanent-allow hashes in the plugin store; may rewrite masked secrets into tool-results copies.

Runs commands

Yes

Runs read-only git subcommands in the project and a bounded find.

Network

None

Introduction

LeakStop is a Claude Code function-hooks mod that inspects what Claude is about to write, print, commit or send, spots API keys, private keys, JWTs and Authorization headers, and asks you to block or allow—or rewrite a plain view of .env to names only. Secrets in tool results, your pasted prompt and large Bash transcripts are masked to a short prefix plus three trailing characters. It stays local: no network, no model calls, and it never reads your environment values. It runs read-only git and a bounded find, and stores permanent allows in the plugin store. Needs Claude Code 2.1.287+. Rule-based help, not a hard guarantee. Preview is illustrative.

Who it is for

  • People worried Claude will write secrets into code, chat or git

Requirements

  • Claude Code ≥ 2.1.287
  • Rule-based assistance, not an absolute anti-leak guarantee.