agentmods.dev

cerebro-code-memory CLAUDE.md

Repository instructions for Cerebro, a Python server that stores a codebase's structure and meaning in a local SQLite database so an AI agent can query it later.

In plain words
What is it for?
Use them before changing Cerebro, particularly when adding tools, editing its Python modules, running tests, or checking that information stays on the local machine.
Why use it?
They explain the real project architecture, dependencies, privacy rule, testing setup, and the established pattern for adding new MCP tools.

Instructions file

6 2mo ago A scan 1,095 tokens original MIT
View the source file marcodavidd020/cerebro-code-memory
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/marcodavidd020/cerebro-code-memory/claude-md
Clone the repo
git clone --depth 1 https://github.com/marcodavidd020/cerebro-code-memory
Per session 1,095 This file is loaded in full into every session.
When invoked 1,095 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 $0.01095 $0.01095
Opus 5 $0.00548 $0.00548
Sonnet 5 $0.00219 $0.00219
Haiku 4.5 $0.00110 $0.00110

Measured 2d ago against content hash b796cfdefdf0, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

cerebro-code-memory 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 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.

CLAUDE.md · 58 lines

How it starts

The opening of the file, as written. The whole thing — 58 lines — stays where its author put it; the contents beside it link to each section on GitHub.

CLAUDE.md — Cerebro

Estándares reales de este proyecto. Cárgalos antes de proponer o escribir código. Si algo aquí contradice al código actual, gana el código: verifícalo y actualiza este archivo.

Qué es

MCP server en Python que cachea el entendimiento de un codebase en un SQLite ("brain") para que sesiones de chat lo consulten en vez de re-leer carpetas. La consigna de todo el proyecto: salida token-barata.

Stack (verificado en pyproject.toml)

  • Python >=3.10. Server con FastMCP (paquete mcp>=1.2.0).
  • Deps núcleo: tree-sitter + tree-sitter-language-pack, networkx, pathspec.
  • Semántica opcional detrás del extra semantic: model2vec, numpy. Sin torch, sin API keys, nada sale de la máquina — invariante de privacidad.
  • Tests: pytest (en .venv). Build: hatchling, paquete en src/cerebro.

Arquitectura (una responsabilidad por módulo, en src/cerebro/)

  • server.py — superficie de tools MCP (@mcp.tool()). Delgada: arma texto compacto y delega la lógica.
  • cli.py — entrypoint CLI unificado (cerebro <subcomando>).
  • config.py — Config.load(); resuelve raíz (CEREBRO_ROOT o git) y db_path.
  • db.py — conexión SQLite + queries de bajo nivel (módulo más central; casi todo depende de él).
  • indexer.py — parseo tree-sitter: símbolos, edges, hashes; reindex/diff.
  • graph.py — grafo de dependencias (dependencies/dependents) sobre networkx.
  • insights.py — impact, cycles, orphans, dead_symbols.
  • callgraph.py — callers/calls (resueltos por nombre).
  • summaries.py / summarizer.py — guardar/leer resúmenes; warming batch vía claude -p headless.
  • embeddings.py — búsqueda semántica model2vec (opcional).
  • notes.py — log de decisiones. gitsync.py — frescura git-aware. tsconfig.py — alias tsconfig/jsconfig.
  • views.py — render de texto (map_text, recall_text). viz.py — HTML del grafo + export Obsidian. docaudit.py — living docs.

Convenciones (observadas en el código, respétalas)

  • from __future__ import annotations; type hints modernos (X | None, no Optional).
  • Toda tool MCP devuelve string compacto. Acota listas con helpers tipo _join_capped y caps (_DEP_CAP, _SYM_CAP). No vuelques datos sin límite: el ahorro de tokens es el producto.
  • Docstrings explican el porqué (economía de tokens, invariantes), no lo obvio.
  • Rutas: siempre normaliza a clave repo-relativa con _resolve_path antes de tocar la DB. Saltarse esto rompe la persistencia entre sesiones.
  • Resúmenes y notas del brain se escriben en inglés (tokeniza más barato), densos, 1-3 frases.
  • Scripts CLI: cada uno expone un main() a nivel módulo (ver [project.scripts]).
  • Tests: fixtures de DB en memoria construyendo edges a mano (ver tests/).

Read the full file on GitHub · 58 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 · 58 lines · 1,095 tokens per session scan A b796cfdefdf0

Subscribe to this mod's changes

cerebro-code-memory CLAUDE.md is an instructions file published in the GitHub repository marcodavidd020/cerebro-code-memory (6 stars, last pushed 2mo ago), licensed MIT. It adds 1,095 tokens to every session, about $0.0055 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.

DatabasesMemory and contextaianthropicclaudeclaude-codecode-intelligencecode-search
Related

Other instructions, from other repositories

repobrain copilot-instructions.md

Copilot instructions for study8677/repobrain, covering github copilot bootstrap instructions and hard rule — query the repobrain hub first.

study8677/repobrain · 229 tokens

repobrain AGENTS.md

AGENTS.md instructions for study8677/repobrain: Beautiful is better than ugly. Explicit is better than implicit. Simple is better than complex. Complex is better than complicated. Flat is better than nested. Sparse is better than dense. Readability counts. Special cases aren't special enough to break the rules.…

study8677/repobrain · 167 tokens

CodeNib AGENTS.md

Instructions for sysevol-ai/CodeNib, covering agents.md, layered objectives, dev commands, git, commit, and pr rules and testing guidance.

sysevol-ai/CodeNib · 1,920 tokens

sverklo CLAUDE.md

Instructions for sverklo/sverklo, covering sverklo development, using sverklo tools, project structure and build & test.

sverklo/sverklo · 406 tokens

sourcebot CLAUDE.md

Claude Code instructions for sourcebot-dev/sourcebot, covering claude code guidelines, database migrations, building packages, backend workloads and execution locks.

sourcebot-dev/sourcebot · 4,674 tokens

codesage AGENTS.md

Instructions for iliaal/codesage, covering codesage, build, sanity check before pushing, crate map and search pipeline.

iliaal/codesage · 6,686 tokens