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/stravu/crystal/claude-mdgit clone --depth 1 https://github.com/stravu/crystalWhat 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 | $0.06391 | $0.06391 |
| Opus 5 | $0.03195 | $0.03195 |
| Sonnet 5 | $0.01278 | $0.01278 |
| Haiku 4.5 | $0.00639 | $0.00639 |
Grade A, and why
crystal 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.
How it starts
The opening of the file, as written. The whole thing — 522 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Crystal - Multi-Session Claude Code Manager
Created by Stravu
Project Overview
Crystal is a fully-implemented Electron desktop application for managing multiple AI code assistant instances (Claude Code and Codex) against a single directory using git worktrees. It provides a streamlined interface for running parallel AI assistant sessions with different approaches to the same problem.
References
Use these reference pages for more information:
- How to invoke Claude Code through the command line as an SDK: https://docs.anthropic.com/en/docs/claude-code/sdk
- How to run multiple Claude Code instances with Git Worktrees: https://docs.anthropic.com/en/docs/claude-code/tutorials#run-parallel-claude-code-sessions-with-git-worktrees
- Adding New CLI Tools: Guide for extending Crystal to support additional CLI tools beyond Claude Code
- Implementing New CLI Agents: Step-by-step instructions for adding new CLI agent tools with code examples and best practices
- Codex Configuration: Configuration guide for Codex CLI integration
Implementation Status: ✅ COMPLETE
All core features have been successfully implemented with significant enhancements beyond the original requirements.
✅ Implemented Features
Core Session Management
- Multi-session support: Run multiple Claude Code instances simultaneously
- Session templates: Create single or multiple sessions with numbered templates
- Session persistence: SQLite database for persistent sessions across restarts
- Session archiving: Archive sessions instead of permanent deletion
- Conversation continuation: Resume conversations with full history context
- Real-time status tracking: initializing, running, waiting, stopped, error
- Automatic session naming: AI-powered session name generation based on prompts
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.
- yesterday First seen · 522 lines · 6,391 tokens per session scan A e00c294ed105
crystal CLAUDE.md is an instructions file published in the GitHub repository stravu/crystal (3,114 stars, last pushed 6mo ago), licensed MIT. It adds 6,391 tokens to every session, about $0.0320 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
claude-code-settings copilot-instructions.md
Instructions for feiskyer/claude-code-settings, covering claude.md, environment setup, required dependencies, configuration and skills.
semantix CLAUDE.md
Instructions for Gnosil/semantix, covering 项目规则（claude code） and 1. 改动须经用户过目；由 claude 提交到分支 + pr，用户 review 后合并.
tree-sitter-analyzer CLAUDE.md
Instructions for aimasteracc/tree-sitter-analyzer, covering ruflo — claude code configuration, rules, test quality rules — locked (2026-06-28), t-1: no new test files for existing plugins (blocker) and t-2: no weak assertions — ratchet-enforced (blocker).
swarm-protocol CLAUDE.md
Instructions for phuryn/swarm-protocol, covering claude.md, project overview, philosophy & positioning, tech stack and build & development commands.
nextjs-skills CLAUDE.md
Instructions for mohamed-hossam1/nextjs-skills: All contribution rules — style, naming, file layout, verification — live in AGENTS.md. Read that first.
Warden-AI CLAUDE.md
Claude Code instructions for rynald0cst0ltziam/Warden-AI, covering warden — context governance and verification layer, session start (important — do this first), layer 1: before starting work — context selection, layer 2: during work — tool output pruning and enforcement hooks (automatic).