toolhive CLAUDE.md

Repository instructions for ToolHive, a Go application that manages MCP servers. MCP is a standard way for AI agents to connect to external tools and data; ToolHive runs those servers in containers and can also manage them through Kubernetes.

In plain words
What is it for?
Use them when building, testing, linting, or extending ToolHive’s command-line tool, Kubernetes operator, proxy runner, or MCP server management.
Why use it?
They explain the system’s components and development commands so an agent can choose the right code and checks when changing the project.

Instructions file

Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

agentmods
npx agentmods add instructions/stacklok/toolhive/claude-md
Clone the repo
git clone --depth 1 https://github.com/stacklok/toolhive
Per session 1,430 This file is loaded in full into every session.
When invoked 1,430 The same file — it is already loaded in full.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5 $0.01430 $0.01430
Opus 5 $0.00715 $0.00715
Sonnet 5 $0.00286 $0.00286
Haiku 4.5 $0.00143 $0.00143

Measured yesterday against content hash da194a7d4327, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

toolhive CLAUDE.md scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured yesterday.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

CLAUDE.md · 107 lines

How it starts

The opening of the file, as written. The whole thing — 107 lines — stays where its author put it; the contents beside it link to each section on GitHub.

CLAUDE.md

This file provides guidance to Claude Code when working with this repository.

Project Overview

ToolHive is a lightweight, secure manager for MCP (Model Context Protocol: https://modelcontextprotocol.io) servers written in Go. It provides a CLI (thv), a Kubernetes operator (thv-operator), and a proxy runner (thv-proxyrunner) for container-based MCP server isolation.

To understand the system, start with the Architecture Documentation. Begin at the Architecture Overview and Core Concepts, then read the component deep-dives relevant to your task — Deployment Modes, Transport Architecture, Kubernetes Operator, Virtual MCP, and more. The architecture index has a full map and by-topic navigation.

Build and Development Commands

task build            # Build the main binary
task install          # Install binary to GOPATH/bin
task lint             # Run linting
task lint-fix         # Fix linting issues (preferred over lint)
task test             # Unit tests (excluding e2e)
task test-e2e         # E2E tests (requires build first)
task test-all         # All tests (unit + e2e)
task test-coverage    # Tests with coverage analysis
task gen              # Generate mocks
task docs             # Generate CLI documentation
task build-image      # Build container image
task build-all-images # Build all container images

IMPORTANT: Always use task commands. Never run go test, go build, or golangci-lint directly -- the Taskfile has correct flags, exclusions, and environment setup that direct commands miss.

Testing: Ginkgo/Gomega for BDD-style tests. Unit tests for pkg/ business logic; E2E tests for CLI commands.

Available Subagents

Agents are in .claude/agents/ and MUST be invoked for tasks matching their expertise:

Read the full file on GitHub · 107 lines

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. yesterday First seen · 107 lines · 1,430 tokens per session scan A da194a7d4327

Subscribe to this mod's changes

toolhive CLAUDE.md is an instructions file published in the GitHub repository stacklok/toolhive (2,056 stars, last pushed today), licensed Apache-2.0. It adds 1,430 tokens to every session, about $0.0072 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.