carapace-sh/carapace-bin / skills/lore
lore
FlaggedUse when working with the Epic Games Lore version control system — a centralized VCS designed for game development with binary-first content-addressed storage. Covers architecture, core concepts, server deployment, CLI and server configuration, workflows, and the full command reference. Triggers on: "lore", "Lore", "Epic Games Lore", "loreserver", "version control system", "VCS", "content-addressed", "immutable store", "mutable store", "BLAKE3", "Merkle tree", "FastCDC", "fragment", "partition", "content-defined chunking", "split-write", "shared store", "revision", "branch", "merge", "cherry-pick", "revert", "bisect", "layer", "link", "lock", "staging", "dirty", "obliterate", "packfile", "LoreError", "config.toml", "cli.toml", "lore://", "Lore server", "Lore CLI", "Lore configuration", "Lore FAQ", "Lore roadmap", "Lore system design".
| File | Kind | Size |
|---|---|---|
| SKILL.md | skill md | 5.9 KB |
| references/cli.md | markdown | 14.0 KB |
| references/config.md | markdown | 8.1 KB |
| references/core-concepts.md | markdown | 11.5 KB |
| references/faq-roadmap.md | markdown | 5.0 KB |
| references/overview.md | markdown | 5.8 KB |
| references/setup.md | markdown | 3.7 KB |
| references/workflow.md | markdown | 5.9 KB |
references/overview.md
View on GitHub# Overview
What Lore is, its design goals, architecture, and how it compares to other version control systems.
> **Source of truth**: <https://epicgames.github.io/lore/explanation/system-design/>
## What Lore Is
Lore is a **centralized, binary-first, MIT-licensed version control system** from Epic Games, designed for projects that combine code with large binary assets (games, entertainment, media). It treats all content as opaque byte streams — text and binary flow through the same primitives.
The project is organized as a **Rust library, server, and CLI**, with publicly specified and versioned data formats and wire protocols. Pre-built binaries ship for Windows, macOS (ARM64), and Linux (x86-64, ARM64).
## Two Subsystems
Lore is structured as two independently usable subsystems:
- **Storage Subsystem**: A partition-based, content-addressed store that deduplicates all content while enforcing strict per-partition access boundaries. Usable on its own for caches, asset pipelines, backup targets.
- **Version Control Subsystem**: Revisions, branches, merges, staging, sync, push, diff, query — built on top of the storage subsystem, exposed through its own public API. Version control is one consumer of the storage API, not a privileged layer.
## Design Goals
- **Binary-first**: All content is opaque byte streams on the hot path. No line-ending translation, no text encoding inference, no clean/smudge filters.
- **Centralized but offline-capable**: Staging, committing, branching, switching, diffing never require a network round trip.
- **Sparse by construction**: Instances materialize only the subset of files asked for, via view filters and lazy fragment fetching.
- **Atomic state**: Every operation completes fully or leaves the repository unchanged.
- **Cryptographically verifiable**: BLAKE3 hashes, hash-chained revision graph.
- **API-first**: A C header is the canonical interface. Language bindings (Rust, Python, JS, C#, Go) wrap the same operations.
- **Multi-tenant safe**: Partition-level isolation with cross-partition dedup but no cross-partition access.
- **Backend-scalable**: Horizontal scaling, hot/warm/cold tiering, edge caching, read replicas.
- **Replaceable backends**: Thin `ImmutableStore` and `MutableStore` trait interfaces. Shipped backends include local file, S3, and DynamoDB.
- **Lifecycle-complete**: Content can be added, deduplicated, and removed (obliterated).
- **Open and free**: MIT license, publicly versioned specs.
## Explicit Non-Goals
- **Peer-to-peer decentralization** (centralized by design)
- **Adversarial-server threat model** (clients hash-validate fragments; trust in the server is bounded to pointers, names, and access decisions)
## Architecture
```
┌──────────────────────────────────────────────────┐
│ Lore CLI │
│ (thin layer over the C API, all commands) │
└──────────────┬───────────────────────────────────┘
│
┌──────────────▼───────────────────────────────────┐
│ Version Control Subsystem │
│ revisions, branches, merges, staging, sync, │
│ push, diff, query, metadata, locks, links, layers│
└──────────────┬───────────────────────────────────┘
│
┌──────────────▼───────────────────────────────────┐
│ Storage Subsystem │
│ Immutable Store (content-addressed, BLAKE3) │
│ Mutable Store (key-value, branch pointers, cas) │
└──────────────┬───────────────────────────────────┘
│
┌──────────────▼───────────────────────────────────┐
│ Replaceable Backends │
│ local (packfiles), S3, DynamoDB, composite, │
│ replicated, remote, plugins │
└──────────────────────────────────────────────────┘
```
## Storage Protocol
One logical command set served on two transports:
- **QUIC** (ALPN `lore-storage/0.4`): Binary, low-overhead, up to 8 bidirectional streams per connection, command pipelining.
- **gRPC** (HTTP/2 + protobuf): Where QUIC is unavailable or existing HTTP infrastructure is preferred.
Commands: Authorize, Get, Put, Query, Verify, Copy (immutable store) + MutableLoad, MutableStore, MutableCas (mutable store). One fragment per command, max 256 KiB.
## How Lore Differs from Other VCS
| Aspect | Git | Perforce | Lore |
|--------|-----|----------|------|
| **Binary assets** | LFS bolted on, second-class | First-class but proprietary | First-class, chunked, deduplicated |
| **Offline** | Fully offline | Server round-trip required | Staging/committing/branching offline |
| **Sparse checkouts** | Sharp edges, offline issues | Server-dependent | Lazy by default, view filters |
| **Deduplication** | File-level only | Limited | Fragment-level, content-defined |
| **Multi-tenant** | Per-repo, no native isolation | Per-depot, limited | Partition-level, access-controlled |
| **License** | GPL | Proprietary | MIT |
| **Protocol** | Custom | Proprietary | Open specification, QUIC/gRPC |Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| writing-skillsobra/superpowersUse when creating new skills, editing existing skills, or verifying skills work before deployment | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| xlsxanthropics/skillsUse this skill any time a spreadsheet file is the primary input or output. This means any task where the user wants to: open, read, edit, or fix an existing .xlsx, .xlsm, .xltx, .csv, or .tsv file (e.g., adding columns, computing formulas, formatting, charting, cleaning messy data); create a new spreadsheet from scratch or from other data sources; or convert between tabular file formats. Trigger especially when the user references a spreadsheet file by name or path — even casually (like "the xlsx in my downloads") — and wants something done to it or produced from it. Also trigger for cleaning or restructuring messy tabular data files (malformed rows, misplaced headers, junk data) into proper spreadsheets. The deliverable must be a spreadsheet file. Do NOT trigger when the primary deliverable is a Word document, HTML report, standalone Python script, database pipeline, or Google Sheets API integration, even if tabular data is involved. | Stars: 198k | Updated | Claude CodeCodex | Verified | Install |
| shipping-and-launchaddyosmani/agent-skillsPrepares production launches. Use when preparing to deploy to production, or when asking what needs to be in place before shipping. Use when you need a pre-launch checklist, when setting up monitoring, when planning a staged rollout, or when you need a rollback strategy. | Stars: 119k | Updated | Claude CodeCodex | Verified | Install |
| childhoodthedotmack/claude-memBabysit a shipped release after merge, publication, or deployment. Verify published packages, production behavior, installed plugins, and late CI or user reports until the observation window is clean; use babysit for PR readiness before release. | Stars: 118k | Updated | Claude CodeCodex | Verified | Install |