graphmind-memory

A set of rules connects the coding agent to GraphMind, a shared memory system that stores project knowledge between sessions. It uses an MCP service to restore context, record actions, and save decisions.

In plain words
What is it for?
Use it to load project memory at the start, record meaningful work and decisions, and save a session summary at the end.
Why use it?
It helps agents recover earlier project context and preserve important facts instead of starting from an empty conversation each time.

Cursor rule

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/kostya-pakhomov/graphmind/graphmind-memory
Clone the repo
git clone --depth 1 https://github.com/kostya-pakhomov/graphmind
Per session 298 This file is loaded in full into every session.
When invoked 298 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 $0.00298 $0.00298
Opus 5 $0.00149 $0.00149
Sonnet 5 $0.00060 $0.00060
Haiku 4.5 $0.00030 $0.00030

Measured 2d ago against content hash c44bb1727f2b, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

graphmind-memory 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.

templates/cursor-automation/rules/graphmind-memory.mdc · 41 lines

What it actually says

GraphMind — память между сессиями

Cursor/Kodik не помнит прошлые чаты. Память — только через MCP graphmind.

Справочник инструментов: docs/TOOLS.md

Старт (до основной работы)

bootstrap_memory({ cwd: "<PROJECT_ROOT>", query: "<суть задачи>", depth: 1 })

cwd — абсолютный путь хоста. Не пропускай.

В процессе — workspace

  • record_action({ summary }) после значимых шагов (≤200 символов).
  • Hook postToolUse: напоминание после 2+ правок (throttle ~45 с).
  • Факт этого репо: propose_new_memory({ level: "L2", node_type: "atom", content }).

Кросс-проектная память (scope: "global")

propose_new_memory({ level: "L2", node_type: "atom", scope: "global", content: "# ..." })
scope: "global" workspace
Hooks, MCP, Docker, CI, конвенции Код и факты только этого приложения

Конец сессии

flush_session_memory({ summary: "итог сессии" })
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 · 41 lines · 298 tokens per session scan A c44bb1727f2b

Subscribe to this mod's changes

graphmind-memory is a cursor rule published in the GitHub repository kostya-pakhomov/graphmind (2 stars, last pushed 1mo ago), licensed MIT. It adds 298 tokens to every session, about $0.0015 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.