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/roberto-mello/lavra/claude-mdgit clone --depth 1 https://github.com/roberto-mello/lavraWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/instructions/roberto-mello/lavra/claude-md)<a href="https://agentmods.dev/instructions/roberto-mello/lavra/claude-md"><img src="https://agentmods.dev/badge/instructions/roberto-mello/lavra/claude-md.svg" alt="Measured on agentmods" height="20"></a>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.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.01801 | $0.01801 |
| Opus 5 | $0.00901 | $0.00901 |
| Sonnet 5 | $0.00360 | $0.00360 |
| Haiku 4.5 | $0.00180 | $0.00180 |
Grade C, and why
lavra CLAUDE.md scanned grade C with 2 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 6d 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.
cat .claude/settings.json | jq . Enumerates other installed skillsmediumAgent snooping
Other skills' SKILL.md files reveal prompts, capabilities and secrets that should be invisible to peers.
ls -la .claude/skills/ How it starts
The opening of the file, as written. The whole thing — 206 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
This is a Claude Code plugin marketplace that provides beads-based persistent memory with compound-engineering's multi-agent workflows. The primary plugin is lavra, located at plugins/lavra/.
The plugin provides:
- Specialized agents across review, research, design, workflow, and docs categories
- Core commands + optional domain-specific commands (in commands/optional/)
- Skills (core + optional: git-worktree, brainstorming, create-agent-skills, agent-browser, dhh-rails-style, etc.)
- 4 hooks (auto-recall, memory-capture, subagent-wrapup, teammate-idle-check)
- 1 MCP server (Context7 for framework documentation)
- Automatic knowledge capture from beads comments (LEARNED/DECISION/FACT/PATTERN/INVESTIGATION/DEVIATION/SKIP)
- Automatic knowledge recall at session start based on current beads
- Workflow configuration via
.lavra/config/lavra.json(toggle research, review, goal verification, parallelism)
Multi-Platform Support
lavra supports OpenCode and Gemini CLI in addition to Claude Code:
OpenCode: Core memory system via native TypeScript plugin at plugins/lavra/opencode/plugin.ts. Commands/agents/skills are Claude Code-specific. OpenCode reads AGENTS.md.
Gemini CLI: Full hook compatibility via gemini-extension.json manifest. Uses same stdin/stdout JSON protocol as Claude Code.
See README.md for detailed setup instructions.
Repository Structure
lavra/ # Marketplace root
├── .claude-plugin/
│ └── marketplace.json # Marketplace catalog
├── plugins/
│ └── lavra/ # Plugin root
│ ├── .claude-plugin/
│ │ └── plugin.json # Plugin manifest (v0.7.0)
│ ├── agents/ # agents (review/, research/, design/, workflow/, docs/)
│ ├── commands/ # core commands + optional/
│ ├── skills/ # skills (core + optional) with supporting files
│ ├── hooks/ # hooks.json, auto-recall.sh, memory-capture.sh, subagent-wrapup.sh, teammate-idle-check.sh, recall.sh
│ ├── opencode/ # OpenCode TypeScript plugin
│ ├── gemini/ # Gemini CLI hook configuration
│ ├── gemini-extension.json # Gemini extension manifest
│ ├── scripts/ # import-plan.sh
│ └── .mcp.json # Context7 MCP server
├── install.sh # Installer (at marketplace root)
├── uninstall.sh # Uninstaller (at marketplace root)
├── CLAUDE.md
└── README.md
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.
- 6d ago First seen · 206 lines · 1,801 tokens per session scan C 53d53bb76202
lavra CLAUDE.md is an instructions file published in the GitHub repository roberto-mello/lavra (50 stars, last pushed 2mo ago), licensed MIT. It adds 1,801 tokens to every session, about $0.0090 per session on Opus 5. A static security scan graded it C with 2 findings (reads agent configuration directories, enumerates other installed skills). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other instructions, from other repositories
beads AGENTS.md
AGENTS.md instructions for gastownhall/beads, covering agent instructions, key sections, project scope, pr safety for agents and visual design anti-patterns.
remind AGENTS.md
Instructions for sandst1/remind, covering remind - development guide for ai agents, project overview, architecture, key abstractions and data models (models.py).
agentor CLAUDE.md
Claude Code instructions for lonetis/agentor, covering agent orchestrator (agentor), architecture, detailed documentation, tech stack and dev commands.
clankerchat CLAUDE.md
Claude Code instructions for simozampa/clankerchat: Preserve concurrent changes, make small Conventional Commits without attribution trailers, and run npm run check before declaring work complete.
yaks AGENTS.md
Instructions for mattwynne/yaks, covering architecture, testing layers, mutation testing, cli design philosophy and development workflow.
Orra-mcp CLAUDE.md
Instructions for bjornj12/Orra-mcp, covering orra mcp development, project, architecture, testing and structure.