brain

Rules for a shared semantic memory service. It loads project guidance at the start of a coding task and records solved problems or technical decisions at the end.

In plain words
What is it for?
Use it to retrieve task guardrails, search stored technical knowledge when needed, and save solutions, rules, pitfalls, or decisions.
Why use it?
It gives the agent known project-specific pitfalls before work starts and preserves useful lessons for later tasks.

Cursor rule for Cursor

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 rules/guilopes29/brain-mcp/brain
Clone the repo
git clone --depth 1 https://github.com/GuiLopes29/brain-mcp

Made for: Cursor.

Per session 722 This file is loaded in full into every session.
When invoked 722 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 $0.00722 $0.00722
Opus 5 $0.00361 $0.00361
Sonnet 5 $0.00144 $0.00144
Haiku 4.5 $0.00072 $0.00072

Measured yesterday against content hash 7a3817fa82a7, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

brain 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 yesterday.

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.

.cursor/rules/brain.mdc · 66 lines

How it starts

The opening of the file, as written. The whole thing — 66 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Brain MCP — Usage Rules

⚡ Task start: load the guardrails (1 cheap call)

BEFORE starting a coding task, call get_guidelines({ project: "<project-name>", source: "cursor" }) once. It returns ≤12 short directives (learned rules and pitfalls) — hundreds of tokens, not thousands. Apply them to the work.

Token-economy rules:

  • Call get_guidelines only once per task, never repeatedly.
  • Use search_knowledge only when you need detail on a specific solution the guardrails mentioned.

Automatic Knowledge Capture

When wrapping up a development session where a problem was solved or a technical decision was made, automatically call add_knowledge with:

  • title: one-line summary of what was solved
  • content: full description — problem, root cause, solution applied, alternatives considered
  • tags: technologies and concepts involved
  • project: name of the repository or project in context
  • source: "cursor"
  • problem: original problem description
  • kind: solution | rule | pitfall | decision
  • directive: ONE imperative, actionable line capturing the lesson (required when kind is rule or pitfall)

Do this without being asked. It's the system's memory.

Quality criteria — only write if at least one of these is true

  • The problem wasn't obvious (required real investigation to find the cause)
  • It cost real time (more than ~10 min of debugging or research)
  • It's a decision another AI would repeat without this context
  • It's a pattern that will show up again in the project

Don't write: trivial fixes, typos, obvious language/framework behavior.

⚠️ NEVER write to the Brain

  • Credentials, tokens, API keys, passwords, connection strings
  • PII of real users (names, emails, national IDs, customer data)
  • Code snippets with identifiable proprietary data from your employer/client
  • Any string that looks like a secret (patterns: AKIA, -----BEGIN, eyJ, Bearer , sk-)

If the content to write contains any of these, omit it or replace it with [REDACTED].

Read the full file on GitHub · 66 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. yesterday First seen · 66 lines · 722 tokens per session scan A 7a3817fa82a7

Subscribe to this mod's changes

brain is a cursor rule published in the GitHub repository GuiLopes29/brain-mcp (0 stars, last pushed 1mo ago), licensed MIT. It adds 722 tokens to every session, about $0.0036 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.