atharvmantri

MemoryGuard — AI skill for Claude Code

AI community

Local-first Context Sync for AI coding agents.

How to install MemoryGuard

This entry records only its repository, not the path inside it, so there is no exact command to give. Open atharvmantri/MemoryGuard and copy the folder into ~/.claude/skills/, or the file into ~/.claude/agents/.

What MemoryGuard does

Local-first Context Sync for AI coding agents. Keeps AGENTS.md, CLAUDE.md, MEMORY.md, and Cursor rules current and secret-safe.

Alternatives in AI

Token Monitor — Local-first desktop widget for tracking token usage, costs, and limits across 32+ AI coding tools—including Cl 1.8k ★
TokenTracker — Local-first AI token usage & cost tracker for 31 coding tools incl 1.4k ★
Project Butler — Project memory system for AI coding assistants (Claude Code, Cursor, Codex): session logs, project wiki, rules 372 ★

README

MemoryGuard

**MemoryGuard keeps AI coding-agent context files current and secret-safe.**

**Stop re-explaining your project to AI coding agents.**

MemoryGuard is a local-first developer CLI that turns durable project facts into generated context files for coding agents. It helps keep `AGENTS.md`, `CLAUDE.md`, `MEMORY.md`, and Cursor rules aligned with the current truth of your project, while redacting secret-looking values before they reach generated context.

Before / After

Before:

"This project uses FastAPI."
"This project uses npm."
"API key: sk-test-..." (fake example)

After MemoryGuard:

"Backend framework: Flask"
"Package manager: pnpm"
"FastAPI was previously used; superseded by Flask."
"Sensitive memory omitted from generated context."

The Problem

AI coding agents forget project decisions between sessions. Humans end up re-explaining the stack, commands, constraints, and gotchas. Worse, old facts can linger in context files after the project changes, and raw transcripts can accidentally contain secrets.

What MemoryGuard Does

Stores approved project memories locally.
Detects obvious supersession, such as FastAPI -> Flask or npm -> pnpm.
Captures candidate memories from local coding-agent transcripts.
Keeps candidates pending until you approve or reject them.
Generates context files for common coding-agent tools.
Redacts common secret-looking values before display, storage, and rendering.
Runs without a cloud service or required external LLM API.

2-Minute Demo (one-line install)

**Windows (PowerShell):**

irm https://raw.githubusercontent.com/atharvmantri/MemoryGuard/main/scripts/install.ps1 | iex

**macOS / Linux:**

curl -fsSL https://raw.githubusercontent.com/atharvmantri/MemoryGuard/main/scripts/install.sh | bash

After the one-line install finishes, run `memoryguard doctor` then `memoryguard demo`. The demo runs in a temporary project and proves transcript capture, pending