crystallized-intelligence AGENTS.md

Repository instructions for Crystallized Intelligence, an open-source framework that organizes specialist knowledge into layers agents can retrieve. The repository contains the framework and tools, while a separate brain directory or repository stores domain knowledge.

In plain words
What is it for?
Use it when contributing to the repository, setting up a brain, or helping an agent retrieve domain guidance through the command-line tools or MCP.
Why use it?
It tells coding agents how to initialize, search, and use that knowledge without loading an entire knowledge collection into context. It also defines how to handle source trust and token limits.

Instructions file for CodexOpenCode

Install

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.

agentmods
npx agentmods add instructions/bcharleson/crystallized-intelligence/agents-md
Clone the repo
git clone --depth 1 https://github.com/bcharleson/crystallized-intelligence

Made for: Codex, OpenCode.

Per session 746 This file is loaded in full into every session.
When invoked 746 The same file — it is already loaded in full.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

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.

ModelPer sessionOnce invoked
Fable 5 $0.00746 $0.00746
Opus 5 $0.00373 $0.00373
Sonnet 5 $0.00149 $0.00149
Haiku 4.5 $0.00075 $0.00075

Measured 2d ago against content hash 43dae471a193, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

crystallized-intelligence 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 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.

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.

AGENTS.md · 71 lines

How it starts

The opening of the file, as written. The whole thing — 71 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Agent Guide — Crystallized Intelligence for Agents

Open-source framework for pre-compiling domain expertise into agent-readable layers. This repo ships tooling and a demo brain; your proprietary content lives in a separate brain repo.

Repo roles

Artifact Location Purpose
Framework This repo (tools/, spec/, schemas) classify, crystallize, verify, retrieve
Brain Your fork's brain dir or sibling repo brain.yaml + corpus/{domain}/
Skill skills/use-crystallized-brain/SKILL.md How agents retrieve without token burn

Quick start for agents

export BRAIN_ROOT=examples/demo-brain   # replace with your brain path
python tools/bin/brain.py bootstrap specialty-coffee

Prefer MCP when configured: tools brain_bootstrap, brain_expand, brain_search, brain_get_document. See docs/AGENT-INTEGRATION.md.

Operating rules

  1. Layer-first — bootstrap (seed + principles) before reading knowledge/ or sources/
  2. Trust-weighted — lower source_tier wins on conflict (0 = first-party)
  3. Budgeted expand — use --max-tokens and --max-tier 3 by default
  4. No corpus dumps — never load all of sources/ into context
  5. Framework only here — no hardcoded company domains or client content in this repo

Copy skills/use-crystallized-brain/SKILL.md into .cursor/skills/ (or your agent's skill path) when working against a real brain.

Fork → initialize → use

# Initialize a new brain (separate directory)
python tools/bin/brain.py init --path ~/my-brain --name "My Brain" --domains "ops"

export BRAIN_ROOT=~/my-brain
# Add corpus files with frontmatter → crystallize → verify
python tools/bin/brain.py crystallize --domain ops --local
python tools/bin/brain.py verify --domain ops

Tools reference

Command Agent use? Notes
brain bootstrap Yes Always first
brain expand Yes Query + token budget
brain search Yes Paths + snippets
brain get Yes Single document
brain domains Yes Discovery
classify Maintainer Validate tiers
crystallize Maintainer Rebuild _crystal/
verify Maintainer Schema + structure
freshness-audit Maintainer Staleness report

Read the full file on GitHub · 71 lines

Changes

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.

  1. 2d ago First seen · 71 lines · 746 tokens per session scan A 43dae471a193

Subscribe to this mod's changes

crystallized-intelligence AGENTS.md is an instructions file published in the GitHub repository bcharleson/crystallized-intelligence (5 stars, last pushed 3mo ago), licensed MIT. It adds 746 tokens to every session, about $0.0037 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.