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/mcs-cli/mcs/claude-mdgit clone --depth 1 https://github.com/mcs-cli/mcsWhat 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 | $0.04415 | $0.04415 |
| Opus 5 | $0.02207 | $0.02207 |
| Sonnet 5 | $0.00883 | $0.00883 |
| Haiku 4.5 | $0.00441 | $0.00441 |
Grade B, and why
mcs CLAUDE.md scanned grade B with 1 finding 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 2d 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.
Reads agent configuration directoriesmediumAgent snooping
.claude/, .codex/, .gemini/ hold keys, settings and other credentials a mod has no legitimate need for.
- `ConfigCommand.swift` — `mcs config list/get/set` for managing user preferences; `set` immediately syncs the SessionStart hook in `~/.claude/settings.json` How it starts
The opening of the file, as written. The whole thing — 210 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.
Project Overview
Swift CLI tool (mcs) that configures Claude Code with MCP servers, plugins, skills, hooks, and settings. Pure pack management engine with zero bundled content — all features come from external tech packs. Distributed via Homebrew.
Commands
# Development
swift build # Build the CLI
swift test # Run tests
swift build -c release --arch arm64 --arch x86_64 # Universal binary
# CLI usage (after install)
mcs sync [path] # Sync project: multi-select packs, compose artifacts (default command)
mcs sync --pack ios # Non-interactive: apply specific packs (repeatable)
mcs sync --all # Apply all registered packs without prompts
mcs sync --dry-run # Preview what would change
mcs sync --customize # Per-pack component selection
mcs sync --global # Sync global scope (MCP servers, brew, plugins to ~/.claude/)
mcs sync --lock # Checkout locked versions from mcs.lock.yaml
mcs update # Fetch latest pack versions and re-apply across every configured scope
mcs update --global # Refresh only the global scope
mcs update --project # Refresh only the current project's scope
mcs update --all-projects # Refresh global + every project tracked in ~/.mcs/projects.yaml (asks confirmation)
mcs update --dry-run # Preview what would change
mcs doctor # Diagnose installation health
mcs doctor --fix # Diagnose and auto-fix issues
mcs doctor --pack ios # Only check a specific pack
mcs doctor --global # Check globally-configured packs only
mcs pack add <source> # Add a tech pack (git URL, GitHub shorthand, or local path)
mcs pack add user/repo # GitHub shorthand → https://github.com/user/repo.git
mcs pack add /path/to/pack # Add a local pack (read in-place, no clone)
mcs pack add <url> --ref <tag> # Add at a specific tag, branch, or commit (git only)
mcs pack add <url> --preview # Preview pack contents without installing
mcs pack remove <name> # Remove an external tech pack
mcs pack remove <name> --force # Remove without confirmation
mcs pack list # List registered external packs
mcs pack update [name] # Refresh pack registry only (low-level fetch; use 'mcs update' for fetch + apply)
mcs pack validate [source] # Validate a tech pack (path, identifier, or current directory)
mcs cleanup # Find and delete backup files
mcs cleanup --force # Delete backups without confirmation
mcs export <dir> # Export current config as a tech pack
mcs export <dir> --global # Export global scope (~/.claude/)
mcs export <dir> --identifier id # Set pack identifier (prompted if omitted)
mcs export <dir> --non-interactive # Include everything without prompts
mcs export <dir> --dry-run # Preview what would be exported
mcs check-updates # Check for pack and CLI updates
mcs check-updates --hook # Run as SessionStart hook (24-hour cooldown, respects config)
mcs check-updates --json # Machine-readable JSON output
mcs config list # Show all settings with current values
mcs config get <key> # Get a specific setting value
mcs config set <key> <value> # Set a configuration value (true/false)
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.
- 2d ago First seen · 210 lines · 4,415 tokens per session scan B 2cdeeae3754d
mcs CLAUDE.md is an instructions file published in the GitHub repository mcs-cli/mcs (93 stars, last pushed 1mo ago), licensed MIT. It adds 4,415 tokens to every session, about $0.0221 per session on Opus 5. A static security scan graded it B with 1 finding (reads agent configuration directories). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other instructions, from other repositories
thinkrail AGENTS.md
Instructions for JetBrains/thinkrail, covering thinkrail, module structure & boundaries (top-priority requirement), engine: pi only, in-process, architecture (three rings) and repo layout.
rn-dev-agent AGENTS.md
Instructions for Lykhoyda/rn-dev-agent, covering repository guide for agents, repository map, editing rules, architecture rules and supported node runtimes.
skill-bill CLAUDE.md
Claude Code instructions for oila-gmbh/skill-bill, a project described as: SkillBill packages your engineering judgment into governed skills, then uses a durable Kotlin runtime to ensure agents actually execute that process consistently, resumably, and with bounded context.
pi-agent-dashboard CLAUDE.md
Instructions for BlackBeltTechnology/pi-agent-dashboard, a project described as: Real-time web dashboard for pi coding-agent sessions. Multi-session view, live chat mirroring, integrated terminal, diff viewer, pi-flows execution, and mobile-first remote control via mDNS or zrok tunnel. Coexists with pi's TUI — doesn't…
mini-swe-agent copilot-instructions.md
Instructions for SWE-agent/mini-swe-agent, covering style guide, test style, bad and good.
openagent CLAUDE.md
Claude Code instructions for the-open-agent/openagent, covering claude.md, commands, architecture, backend (go / beego) and frontend (react).