memory-merge

Resolve conflicts and consolidate duplicate memory files in .claude/memory/.

2|Updated Apr 10, 2026
One-click install
npx skills add https://github.com/t0lab/harness-kit --skill memory-merge
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: memory-merge
Source: https://github.com/t0lab/harness-kit/tree/main/.agents/skills/memory-merge
Command: npx skills add https://github.com/t0lab/harness-kit --skill memory-merge

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This Skill resolves conflicts and consolidates duplicate or overlapping team memory files stored in .claude/memory/, preventing data loss while reducing redundancy and improving scannability of project memory.

Core Features & Use Cases

  • Detects conflicted files and stray merge markers, identifies overlapping topics across deep files, and flags oversized hot caches like project.md or reference.md.
  • Classifies conflict types (additive, semantic duplicate, conflicting content, duplicate files) and prescribes curated resolutions that preserve authorship, timestamps, and provenance.
  • Proposes a clear diff plan for user approval, applies merges with metadata updates (last-updated, supersedes), safely removes dropped files via git, and verifies no conflict markers remain.
  • Use cases: resolve memory conflicts after a git pull, consolidate duplicate topic files (e.g., auth.md vs auth-decision.md), and prune oversized project.md to keep the hot cache within budget.

Quick Start

Ask the agent to detect and propose a safe merge plan for conflicts in .claude/memory/ and then apply the approved changes.

Frequently Asked Questions about memory-merge

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I resolve merge conflicts in .claude/memory/ files after a git pull?

To resolve merge conflicts in .claude/memory/ files, you can detect conflicted files and stray merge markers, classify the conflict type, and propose a curated diff plan that preserves authorship and timestamps for explicit approval before committing.

What is the best way to consolidate duplicate memory files covering overlapping topics?

Consolidating duplicate memory files involves classifying overlaps as additive, semantic duplicate, or conflicting content, then prescribing curated resolutions that update supersedes metadata and safely remove dropped files via git as a single commit.

How do I prune an oversized project.md or reference.md memory file?

Pruning an oversized project.md or reference.md memory file requires flagging files that exceed the hot-cache budget, proposing a diff plan to reduce redundancy, and applying merges with updated last-updated metadata after user approval.

Does merging memory files preserve original authorship and timestamps?

Yes, merging memory files preserves original authorship, timestamps, and provenance while proposing diffs for explicit user approval, ensuring no data loss occurs during the consolidation of overlapping or duplicate content.

When should I perform a memory cleanup and conflict resolution on my repository?

You should perform memory cleanup when git pull introduces conflict markers, when two files cover overlapping topics, when project.md or reference.md grows beyond the hot-cache budget, or when explicitly requesting memory consolidation.