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/mrboups/xbrain/claude-mdgit clone --depth 1 https://github.com/mrboups/xbrainWrote 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/mrboups/xbrain/claude-md)<a href="https://agentmods.dev/instructions/mrboups/xbrain/claude-md"><img src="https://agentmods.dev/badge/instructions/mrboups/xbrain/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 | $0.05300 | $0.05300 |
| Opus 5 | $0.02650 | $0.02650 |
| Sonnet 5 | $0.01060 | $0.01060 |
| Haiku 4.5 | $0.00530 | $0.00530 |
Grade A, and why
xbrain 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 today.
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 — 218 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 status
In production. 27 phases have shipped (Phase 1 on 2026-05-03 through Phase 27 on 2026-08-01), deployed on a GCP VM via Docker Compose. Concretely, as of 2026-08-13:
apps/holds 19 services plus one spike;infrastructure/docker-compose.ymldefines 34 services across four opt-in profiles.apps/memory-apiis the single write path — FastAPI + SQLAlchemy, alembic head0034_team_message_private_lane..planning/phases/carries 28 directories — 01 → 27, plus the inserted 03.5.
Vision (one line): xbrain is a collective persistent-memory system for humans + agents organised by team — not a chatbot workspace. Differentiator is the memory + truth-level + team-scope layer, not the frontend.
.planning/PROJECT.md, REQUIREMENTS.md, ROADMAP.md, and STATE.md are the authoritative planning sources — prefer them over this file. When they disagree with the code, the code wins and the planning file is stale; say so rather than working from it.
Workflow: GSD is the build system
Development flows through GSD slash commands, not ad-hoc edits. The hooks in .claude/settings.json (workflow guard, phase boundary, commit validation, read-before-edit guard) enforce this — they fire on Write/Edit/Bash tool calls. Bypassing them defeats the point.
Entry points (run in order on a fresh project):
/gsd:new-project— initialises.planning/(PROJECT.md, REQUIREMENTS.md, ROADMAP.md, STATE.md). Use--auto @docs/idea.mdto skip the wizard when the brief is already written./gsd:plan-phase <N>— produces a sub-plan for phase N before any code is touched./gsd:execute-phase <N>— runs the sub-plan with verification gates and atomic commits per task.
/gsd:help lists the full 65-command surface. Common adjuncts: /gsd:resume-work, /gsd:progress, /gsd:debug, /gsd:review, /gsd:undo.
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.
- today First seen · 218 lines · 5,300 tokens per session scan A 29efc789e12a
xbrain CLAUDE.md is an instructions file published in the GitHub repository mrboups/xbrain (2 stars, last pushed 20d ago), licensed MIT. It adds 5,300 tokens to every session, about $0.0265 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 2026-09-03.
Other instructions, from other repositories
trinity AGENTS.md
AGENTS.md instructions for trinity-tick/trinity, covering agents.md — trinity memory, trinity 记忆层实时快照（生成于 2026-09-02 03:21:12，snapshot 任务自动刷新）, 活跃目标（active goals）, 最近会话（recent sessions） and 1. trinity 是什么.
engrava-mcp CLAUDE.md
Claude Code instructions for sovantica/engrava-mcp, covering claude.md — engrava-mcp, context, repository-specific rules, code quality — zero-compromise bar and run locally before opening a pr.
inite-brain-service AGENTS.md
AGENTS.md instructions for inite-ai/inite-brain-service, covering agents.md — using brain as your memory over mcp, connect, tool surface, memory semantics you must know and common mistakes.
pensyve copilot-instructions.md
Copilot instructions for major7apps/pensyve, covering pensyve working-memory substrate for vs code copilot, memory reflex rule, "substantive answer" means, "lesson lands" means and memory classification.
engrava-mcp copilot-instructions.md
Copilot instructions for sovantica/engrava-mcp, covering copilot instructions — engrava-mcp, commit messages, branch names, code style and what to avoid.
pensyve AGENTS.md
AGENTS.md instructions for major7apps/pensyve, covering pensyve — agents.md, what this repo is, start here, directory map and build & dev commands.