akasecurity

Claude Tools — Development skill for Claude Code

Development community

The clean slate a fresh Claude Code should start from — clean context, locked credentials, guarded egress.

How to install Claude Tools

This entry records only its repository, not the path inside it, so there is no exact command to give. Open akasecurity/claude-tools and copy the folder into ~/.claude/skills/, or the file into ~/.claude/agents/.

What Claude Tools does

The clean slate a fresh Claude Code should start from — clean context, locked credentials, guarded egress. Isolated profile, pick-and-choose.

Alternatives in Development

  • Claude Supermemory — Persistent memory across sessions and projects using Supermemory 2.3k ★
  • Magnitude — Open source inference server your agent sets up for you 1.5k ★
  • Add Transitions — Add professional scene transitions to your Remotion video 113 ★

README

aka-claude-tools

aka-claude-tools: clean context, locked doors, guarded exits for Claude Code. MIT · needs jq + bun.

![version](https://img.shields.io/github/v/tag/akasecurity/claude-tools?label=version&color=blue) ![license](https://img.shields.io/badge/license-MIT-green)

**Make Claude Code safer to use.** Clean context, locked-down credentials, guarded egress. The security defaults Claude Code doesn't ship with, layered onto a profile of its own in a few minutes.

New to this? Hand the repo to Claude and it sets you up. Comfortable in a terminal? Read every hook first. It's all plain shell and TypeScript, MIT, and the guards scan locally: nothing is uploaded to run them.

Also known as `aka-claude-tools` (the npm package, Homebrew formula, and CLI name). Repo: `akasecurity/claude-tools`. The guard-hooks plugin installs as `aka-claude-tools@akasecurity`.

From [akasecurity](https://akasecurity.io) · MIT · needs `jq` + `bun`.


What it guards against

A coding agent runs shell commands and reaches the network on your behalf. This kit adds the guardrails for the obvious foot-guns:

  • Reading your credentials (SSH keys, cloud tokens, .env files, keychains) → denied.
  • curl … | bash and friends (piping a web script straight into your shell) → blocked.
  • Editing your shell startup files (a common way things quietly persist) → blocked.
  • Secrets leaving in a web request → matched on your machine and blocked.
  • Secrets already in the model's context (a file it read, a page it fetched, a search result, an MCP response) → redacted before the model sees them; fetched or MCP content carrying a prompt-injection phrase like "ignore previous instructions" is flagged too.
  • Context filling with noise (chatty command output) → summarized before it reaches the model.
  • A kit file quietly edited, or a security setting reverted → flagged at the next session start,