vanzan01

cursor-memory-bank: Structure AI-Assisted Development Workflows

cursor-memory-bank adds a persistent task memory and phased workflow to Cursor, guiding projects from initialization and planning through implementation and review. It suits developers who want repeatable processes and durable context for AI-assisted coding.

View on GitHub Compare
Stars
3.1k
Forks
437
Stars, last 30 days
-5

Our take

Health score

31/100

How it is scored
Activity
31
Community
50
Issues
21
Pull requests
20

-5 stars in the last 30 days

Project history

Daily GitHub snapshots since Oct 7, 2026. History covers stars and forks; health signals are not tracked historically.

Selected snapshot

Stars

3,054

Forks

428

Change since Nov 15, 2026: 0 stars·0 forks

Overview

cursor-memory-bank is a documentation-driven workflow framework for Cursor. Its commands guide AI-assisted development through task setup, planning, design decisions, implementation, reflection, and archiving, while shared Memory Bank files preserve project and task context between phases.

It is most useful when coding tasks benefit from explicit plans and recorded decisions rather than one-off chat context. The workflow adapts its steps and documentation to task complexity, so small fixes can take a shorter path than larger features.

Key Features

  • Six Cursor commands cover initialization, planning, design exploration, implementation, reflection, and archiving.
  • Shared Markdown files track project context, active tasks, progress, and completed work.
  • Complexity levels adjust the workflow and documentation to the task.
  • Hierarchical and lazy-loaded rules aim to limit how much guidance is loaded at once.
  • Planning can flag components for a separate design exploration phase.
  • Visual process maps and platform-aware command guidance support the workflow.

Use Cases

  • Cursor users handling multi-step features who want plans, design choices, and implementation notes kept together.
  • Developers working on quick fixes who want a lightweight path that still records outcomes.
  • Small teams or individuals who need a repeatable process for carrying project context between AI-assisted coding sessions.
  • Developers maintaining complex changes who want task decisions and lessons preserved for later reference.

Getting Started

Requires Cursor 2.0 or higher. Clone the repository into your project directory:

git clone https://github.com/vanzan01/cursor-memory-bank.git

Start in Cursor with /van and a task description. See the README for setup and workflow details.

Alternatives

  • TaskSync: TaskSync organizes AI coding work with task queues and feedback loops, rather than Cursor-specific persistent memory and phased project guidance.
  • context-engineering-intro: This template prepares Claude Code with project rules and feature plans, but does not provide Cursor-specific persistent task memory.
  • code-session-memory: code-session-memory indexes coding conversations for semantic search across tools, rather than guiding work through a phased project workflow.
ProjectLanguageLicenseStarsStatus
cursor-memory-bank––3.1kNot checked yet
TaskSyncTypeScriptMIT634Archived
context-engineering-introPythonMIT13.9kQuiet
code-session-memoryTypeScript–20Active

Considerations

  • The README describes v0.8 as a beta version under ongoing development.
  • The current workflow requires Cursor 2.0 or higher; legacy custom modes are deprecated in favor of Cursor commands.
  • The maintainer describes this as a personal project and does not actively maintain an issue tracker or collect feedback.
  • The README recommends Claude 4 Sonnet or Claude 4 Opus for best results, particularly for the /creative command. These are recommendations, not stated prerequisites.

Found this useful?

Share it with someone who would like cursor-memory-bank.

Comparisons