Where is the memory data stored?

All memories are stored locally on your machine in a SQLite database located at ~/.claude/pact-memory/memory.db, ensuring your data remains private.

How does graph-enhanced retrieval help my workflow?

It automatically links memories to the files you are working on, so when you open a file, the agent can instantly recall previous decisions or issues related to that specific component.

How does semantic search work in this skill?

It uses local vector embeddings to find relevant memories based on their meaning and context rather than just matching keywords, making retrieval more accurate.

What is the PACT Memory skill?

It is a persistent storage system that allows Claude Code agents to remember project goals, technical decisions, and lessons across different coding sessions.

Do I need an internet connection for the embeddings to work?

No, the skill is designed to use local GGUF models for embeddings, allowing for offline semantic search and better privacy for your code context.

PACT Memory Agent

Name: PACT Memory Agent
Author: ProfSynapse

byProfSynapse

•

생산성 및 워크플로

Provides persistent, cross-session memory for AI coding agents to track project context, architectural decisions, and lessons learned.

소개

PACT Memory is a sophisticated persistent storage system designed for AI-driven development workflows. It allows agents to retain deep context across sessions by storing and retrieving goals, task progress, architectural decisions with their rationales, and actionable lessons learned. Utilizing semantic search with local embeddings and graph-enhanced retrieval, it ensures that previous knowledge is surfaced when relevant files are modified, preventing repeated mistakes and maintaining continuity in complex software projects without needing to manually re-explain context in every session.

주요 기능

Seamless integration with PACT framework phases for systematic project management.
Graph-enhanced retrieval that boosts memory relevance based on file-to-file relationships.
Semantic search across past sessions using local embeddings for natural language retrieval.
36 GitHub stars
Persistent local storage in SQLite for privacy and high-performance access.
Rich memory objects for tracking specific goals, technical entities, and project history.

사용 사례

Recalling the rationale behind complex architectural decisions made in previous weeks.
Accessing specific lessons learned from past bug fixes to avoid repeating the same errors.
Maintaining project context when switching between multiple feature branches or modules.

소개

주요 기능

Seamless integration with PACT framework phases for systematic project management.
Graph-enhanced retrieval that boosts memory relevance based on file-to-file relationships.
Semantic search across past sessions using local embeddings for natural language retrieval.
36 GitHub stars
Persistent local storage in SQLite for privacy and high-performance access.
Rich memory objects for tracking specific goals, technical entities, and project history.

사용 사례

Recalling the rationale behind complex architectural decisions made in previous weeks.
Accessing specific lessons learned from past bug fixes to avoid repeating the same errors.
Maintaining project context when switching between multiple feature branches or modules.