Agent Kit & Repo Audit

Agent Kit & Repo Audit helps developers using AI coding assistants (Claude Code, Codex, Cursor, or GitHub Copilot) make sure the assistant’s instructions actually reflect how their repository works.

It includes a one-time Agent Kit download and a separate Repo Audit. The Kit provides a shared AGENTS.md and CLAUDE.md plus tool-specific rules and workflow prompts. It also adds guard hooks that deny secret-looking paths and prompt before destructive shell commands, helping keep agents inside the boundaries you define. The workflow prompts cover common cycles like planning, bug fixing, refactoring, review, test backfilling, and handoff. A readiness scanner/checklist helps you validate the installed setup and permissions.

The Repo Audit reads your repository and checks the instruction claims against the code and configuration. The audit resolves conflicts and load order across guidance files, verifies paths, commands, scripts, packages, and tool rules against the repository tree, and produces a prioritized list of concrete fixes with evidence. It performs a static review only—your code is never run.

Pricing is one-time and in EUR (no VAT charged under § 19 UStG). Choose Agent Kit for €39, Repo Audit for €149, or Kit + Audit for €169. The audit covers one public repository, up to 10 instruction files and 1,200 instruction lines; for larger repositories, you’re contacted about what will be covered.

Which AI coding assistants does Agent Kit & Repo Audit support?

It’s designed for developers using Claude Code, Codex, Cursor, or GitHub Copilot.

What does the Agent Kit include?

It’s a one-time download containing shared instruction files (AGENTS.md and CLAUDE.md), tool-specific rules/workflows for editing, testing, security, git, and frontend/backend (including Go and TypeScript), a guard hook for Claude Code and Cursor, and workflow prompts plus an agent-readiness scanner/checklist.

What does the Repo Audit check, and does it run my code?

The audit checks your instruction claims against your repository tree—paths, commands, scripts, packages, and tool rules—while resolving conflicts and clarifying load order. It never runs your code (static review only).

How scoped is the audit for a repository?

It covers one public repository with up to 10 instruction files and 1,200 instruction lines. Larger repositories are reviewed with you on what will be covered.

What are the costs and payment model?

There are one-time EUR prices (no VAT charged under § 19 UStG): Agent Kit €39, Repo Audit €149, and Kit + Audit €169.

Can I get feedback before purchasing an audit?

You can paste a file into the free AGENTS.md checker to get a first look at findings around scope, permissions, and structure.