memory-bank

A set of project-documentation rules that keeps key information in a group of Markdown files. These files describe the project, its goals, design, technology, current work, and progress so an AI assistant can restore context after a session reset.

In plain words
What is it for?
Use it to maintain project briefs, technical notes, architecture explanations, current-task notes, progress records, and documentation for source files.
Why use it?
It reduces the risk of losing important project context between sessions. It also gives the assistant a required place to record decisions and changes.

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/barisercan/cursorrules/memory-bank
Clone the repo
git clone --depth 1 https://github.com/barisercan/cursorrules
Per session 458 This file is loaded in full into every session.
When invoked 458 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 $0.00458 $0.00458
Opus 5 $0.00229 $0.00229
Sonnet 5 $0.00092 $0.00092
Haiku 4.5 $0.00046 $0.00046

Measured 2d ago against content hash 6b1de9e59e8c, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

memory-bank 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.

memory-bank.mdc · 67 lines

What it actually says


description: Maintain concise project documentation for context continuity across memory resets globs: "**/*" alwaysApply: true

Memory Bank System

Memory resets between sessions. MUST read ALL memory bank files at start of EVERY task.

actions:

  • type: enforce message: | MEMORY BANK PROTOCOL

    CRITICAL: Memory bank files are the ONLY resource available after session reset. Agent must understand EVERYTHING just by reading these files. No other context exists.

    REQUIRED FILES (memory-bank/):

    1. projectbrief.md - Core requirements and scope
    2. productContext.md - What this solves and user goals
    3. activeContext.md - Current focus and next steps
    4. systemPatterns.md - Architecture and key decisions
    5. techContext.md - Technologies and constraints
    6. progress.md - Status and known issues
    7. view.md - Architecture diagrams
    8. [filename].md - For each code file explanation

    WORKFLOW:

    • Start: Read ALL memory bank files
    • Work: Update documentation after changes
    • Cleanup: Remove outdated info, keep minimal

    UPDATE TRIGGERS:

    • New project patterns discovered
    • After significant changes
    • User says "update memory bank"
    • Context needs clarification

    BREVITY RULE:

    • Each file as brief as possible
    • 1 sentence better than 100
    • Clean out outdated info
    • No unnecessary details

examples:

  • input: "update memory bank" output: "Review ALL files, update current state, remove outdated info"

metadata: priority: critical version: 2.0 enforcement: strict

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 · 67 lines · 458 tokens per session scan A 6b1de9e59e8c

Subscribe to this mod's changes

memory-bank is a cursor rule published in the GitHub repository barisercan/cursorrules (35 stars, last pushed 1y ago), licensed MIT. It adds 458 tokens to every session, about $0.0023 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 2026-08-30.