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.
npx agentmods add instructions/giantswarm/muster/claude-mdgit clone --depth 1 https://github.com/giantswarm/musterWhat 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.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5 | $0.00953 | $0.00953 |
| Opus 5 | $0.00477 | $0.00477 |
| Sonnet 5 | $0.00191 | $0.00191 |
| Haiku 4.5 | $0.00095 | $0.00095 |
Grade A, and why
muster 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 3d ago.
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.
How it starts
The opening of the file, as written. The whole thing — 86 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 (claude.ai/code) when working with code in this repository.
What is Muster
Muster is a universal control plane and meta-MCP (Model Context Protocol) server aggregator for platform engineers and AI agents. It manages multiple backend MCP servers, provides intelligent tool discovery/filtering, includes a mcp-native workflow engine, and exposes its own functionality through mcp tools as well.
Build & Development Commands
# Build and install
make build # Build local binary
go install # Install binary (use after every code change)
# Testing
make test # Unit tests with race detector
muster test --scenario <name> --verbose # Run a single BDD scenario
muster test --scenario <name> --verbose --debug # With debug output
muster test --parallel 50 --base-port 30000 # Run all scenarios
# Code quality (run before every commit)
goimports -w . && go fmt ./... # Format code
make lint # golangci-lint with gosec and goconst
make vet # go vet
# Helm
make helm-lint # Lint Helm chart
Architecture
Service Locator Pattern (mandatory)
All inter-package communication goes through internal/api. This is the most critical architectural rule:
internal/apidefines handler interfaces inhandlers.goand depends on no other internal package- Each service package creates an
api_adapter.goimplementing the handler interface - Adapters register themselves via
api.RegisterXxx()functions - Consumers retrieve handlers via
api.GetXxx()— never import service packages directly - Anti-pattern: Never import
workflowormcpserverpackages directly
Key Packages
cmd/— CLI commands (serve, agent, auth, test, create, get, list)internal/api/— Central service locator with handler interfaces and type definitionsinternal/app/— Application bootstrap, lifecycle, config loading, service registrationinternal/aggregator/— MCP server aggregation, registry, transports (SSE, stdio, streamable-http)internal/agent/— MCP client, interactive REPL, MCP server mode for AI assistantsinternal/mcpserver/— MCP server process management and health checkinginternal/workflow/— Workflow definitions, execution engine, templatinginternal/orchestrator/— Service lifecycle orchestration and state machineinternal/testing/— BDD test framework, scenarios inscenarios/*.yamlpkg/— Public packages (Kubernetes APIs, auth, logging, OAuth)
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.
- 3d ago First seen · 86 lines · 953 tokens per session scan A dd13ce650ad9
muster CLAUDE.md is an instructions file published in the GitHub repository giantswarm/muster (28 stars, last pushed 3d ago), licensed Apache-2.0. It adds 953 tokens to every session, about $0.0048 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.
Other instructions, from other repositories
agent CLAUDE.md
Instructions for 1mcp-app/agent, covering claude.md, repository information, development commands, setup and build & run.
MCPbundler AGENTS.md
Instructions for eugenepyvovarov/MCPbundler, covering repository guidelines, project structure & module organization, documentation (obsidian), build, test, and development commands and headless stdio environment flags.
java-sdk AGENTS.md
AGENTS.md instructions for modelcontextprotocol/java-sdk, covering mcp java sdk, modules, prerequisites, build & test and evolving mcpschema records.
mcp-dotnet-samples AGENTS.md
Instructions for microsoft/mcp-dotnet-samples, covering agents.md - mcp .net samples, project overview, project structure and architecture, repository organization and sample components.
Embody AGENTS.md
Instructions for dylanroscover/Embody, covering embody + envoy - ai instructions, critical rules, td python rules, network layout rules and mcp server safety rules.
one-mcp AGENTS.md
Instructions for burugo/one-mcp, covering repository guidelines, project structure & module organization, build, test, and development commands, coding style & naming conventions and testing guidelines.