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/xjli1972/book-brain/agents-mdgit clone --depth 1 https://github.com/xjli1972/book-brainWrote 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/xjli1972/book-brain/agents-md)<a href="https://agentmods.dev/instructions/xjli1972/book-brain/agents-md"><img src="https://agentmods.dev/badge/instructions/xjli1972/book-brain/agents-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 | $0.01375 | $0.01375 |
| Opus 5 | $0.00687 | $0.00687 |
| Sonnet 5 | $0.00275 | $0.00275 |
| Haiku 4.5 | $0.00137 | $0.00137 |
Grade A, and why
book-brain AGENTS.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 5d 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 — 124 lines — stays where its author put it; the contents beside it link to each section on GitHub.
AGENTS.md — operating the book-brain
Use this file as the agent contract for
book-brain. It mirrors GBrain's AGENTS.md / CLAUDE.md split: tell the agent what the brain is, how to grow it, how to query it, and where the trust boundaries are.
What the brain is
book-brain is a personal RAG over open-source / free books. Each book
splits into per-section markdown pages on disk; SQLite + sqlite-vec is
the index; an MCP server exposes search and read tools.
- Source of truth (runtime data):
$BOOKBRAIN_ROOT/books/<slug>/and$BOOKBRAIN_ROOT/concepts/. Defaults to~/book-brain/ifBOOKBRAIN_ROOTis unset. - Index:
$BOOKBRAIN_ROOT/db.sqlite. - Embedding model:
nomic-embed-textvia Ollama athttp://localhost:11434. - Repo (read-only code): wherever the user cloned this repository. The repo is the engine; the brain is the runtime data directory. Keep them separate.
How to grow the brain
To add a new book (run from the repo, against your runtime brain):
npx tsx src/cli.ts ingest /path/to/book.epub
npx tsx src/cli.ts embed
Both commands are idempotent on the page content hash. Re-running ingest
on an updated EPUB only re-embeds chunks whose body actually changed.
When choosing a book to add:
- Verify the license. Public domain, CC, or "free online" with the author's blessing — yes. DRM'd or strictly commercial — no.
- Prefer EPUB. PDFs need a separate
pdf2epubstep; HTML directories need a custom adapter (out of scope for v0). - Sanity-check the slug.
bookbrain ingest --slug <override>if the auto-derived one collides or reads badly.
How to query
For agents, prefer the MCP tools (start with search, drill into read_page):
| Tool | When to use |
|---|---|
search |
Always start here. Returns ranked passages with vec_rank, fts_rank, backlinks, score, snippet, slug. |
read_page |
Once you've picked a slug from search, get the full markdown. |
concept_neighborhood |
Exploring a topic. Returns mentioning pages + co-occurring concepts (depth=2). |
list_books |
Before constraining a search to one book. |
stats |
Sanity check that the brain is populated and embedded. |
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.
- 5d ago First seen · 124 lines · 1,375 tokens per session scan A 42661c03f2d1
book-brain AGENTS.md is an instructions file published in the GitHub repository xjli1972/book-brain (5 stars, last pushed 4mo ago), licensed MIT. It adds 1,375 tokens to every session, about $0.0069 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 2026-08-31.
Other instructions, from other repositories
Binder AGENTS.md
Instructions for mpazik/Binder, covering binder, tech stack, monorepo structure, testing and development.
knowledge-base CLAUDE.md
Claude Code instructions for riemannulus/knowledge-base, covering knowledge-base 개발 가이드, 명령어, 아키텍처 (데이터 흐름), 깨뜨리기 쉬운 불변 원칙 and 확장 포인트.
Soroka AGENTS.md
Instructions for AndyShaman/Soroka, covering agents.md — deployment protocol for ai agents, required values from the user, deployment, hand-off and diagnostics.
auto-knowledge-base CLAUDE.md
Claude Code instructions for guyeyouhun/auto-knowledge-base, covering 自主知识库 (auto knowledge base), 当前状态, 核心设计原则, 知识类型 and 信任级别.
fortemi CLAUDE.md
Claude Code instructions for Fortemi/fortemi, covering matric memory, ci/cd, gitea actions workflows, monitoring builds and check recent runs via mcp.
Soroka CLAUDE.md
Instructions for AndyShaman/Soroka, covering soroka and session-start rule.