Loaf — Development skill for Claude Code
🍞 Loaf – An opinionated agent harness/framework – Built primarily for Claude Code, with support for Codex, OpenCode, and Cursor.
How to install Loaf
This entry records only its repository, not the path inside it, so there is no
exact command to give. Open levifig/loaf and copy the folder into
~/.claude/skills/, or the file into ~/.claude/agents/.
What Loaf does
🍞 Loaf – An opinionated agent harness/framework – Built primarily for Claude Code, with support for Codex, OpenCode, and Cursor.
Alternatives in Development
- IPolloWork — Enterprise-grade, local-first Agent Workbench for people and agent teams 4.9k ★
- Petdex — A public gallery of animated pets for Codex, Claude Code, DeepSeek Harness, Hermes, OpenCode, Gemini CLI, and 4k ★
- Ralphy — My Ralph Wiggum setup, an autonomous bash script that runs Claude Code, Codex, OpenCode, Cursor agent, Qwen & 2.8k ★
README
🍞 Loaf
"Why have just a slice when you can get the whole loaf?"
Loaf brings shared skills, a native Go CLI, and tracker-native workflows to Claude Code, OpenCode, Cursor, Codex, and Amp. Skills guide judgment, the CLI handles deterministic checks and private continuity, and your tracker owns shared work.
This README describes the current source tree. Until its next release is published, installation from GitHub Releases or Homebrew may provide an earlier version; see the [published releases](https://github.com/levifig/loaf/releases) and [unreleased changes](CHANGELOG.md#unreleased).
Why Loaf?
**Portable knowledge** — Skills cover workflows, engineering standards, and language expertise. Build once, deploy to supported AI coding tools without rewriting anything.
**Private continuity** — A project-scoped SQLite journal records decisions, discoveries, and optional wrap summaries across conversations and worktrees. There is no session entity to open or close. Startup context is derived from the journal and Git; useful resumption depends on recording meaningful context, not retaining a second work queue.
**Tracker-native workflow** — Pitch → shape → implement → ship → release keeps discovery, selected work, review, and publication distinct. GitHub and Linear provider skills use connections already available in your harness; Loaf never synchronizes a parallel local issue database, and does not configure their credentials. On GitHub, the selected board's Status distinguishes noncommittal Backlog from selected Todo work.
**Profile-based agents** — Functional profiles are defined by tool access, not job titles. A Smith with `python-development` skills becomes a backend engineer; the same Smith with `infrastructure-management` becomes a DevOps engineer. Skills determine what an agent knows; the profile determines what it can touch.
**Living architecture** — Keep current design and rationale in topic-based `docs/architecture/` documents. Architecture and Refle
Related Skills
Tokrs
A command-line interface (CLI) for analyzing token usage in local logs, supporting Claude, Codex, OpenCode, Ki
Cc Proxy
Use Claude Code with your OpenAI Codex, Kimi, Grok, OpenCode Go or Cursor Agent subscription. Local proxy with
Ds Harness Remote
基于 DeepSeek Harness 插件机制构建的远程访问方案,通过低延迟、端到端加密的 P2P 网络,支持使用 PC、移动端 和 Web 随时访问本地设备上的DSH 、CodeX、Cursor、Claude Cod
Oh My Agent
Portable multi-agent harness for .agents-based skills, workflows, and standards-aware agent teams across Antig
Agentrules Architect
AGENTS.md/CLAUDE.md generator and ExecPlan harness for Codex, Claude Code, Cursor, Antigravity, OpenCode and o
Opencostudio Releases
Not another harness. One studio for the coding agents you already use — Codex, Claude Code, Cursor, Antigravit
Related Agents
Async Code
Use Claude Code or CodeX CLI to perform multiple tasks in parallel with a Codex-style UI, functioning as a per
Monk Installer
Install, upgrade, and troubleshoot monk-agent plus Monk CLI/daemon for Claude, Codex, and Cursor users.
Plugin Host Parity Checker
Checks that the shared skills/ tree, the three host manifests (.claude-plugin/, .cursor-plugin/, .codex-plugin