Byori — Development skill for Claude Code
The multi-agent coding workspace that remembers your project — run Claude Code and Codex side by side over a shared ByoriDB graph memory.
How to install Byori
This entry records only its repository, not the path inside it, so there is no
exact command to give. Open byoridb/byori and copy the folder into
~/.claude/skills/, or the file into ~/.claude/agents/.
What Byori does
The multi-agent coding workspace that remembers your project — run Claude Code and Codex side by side over a shared ByoriDB graph memory.
Alternatives in Development
- IPolloWork — Enterprise-grade, local-first Agent Workbench for people and agent teams 4.9k ★
- Understand Anything — Claude Code skills that turn any codebase into an interactive knowledge graph you can explore, search, and ask 2.5k ★
- Zeroshot — Runs coding agents as a graph: one agent implements, independent agents review, failures go to repair, and not 1.9k ★
README
**English** | [한국어](docs/ko/README.md)
Byori
**The memory of your software project.**
Git remembers what changed. Byori remembers why.
Every coding agent forgets. Your project shouldn't.
$ byori init # reads the repository's own history
analyzed
11,814 commits · 1,858 tracked files · 1,042 pull requests
discovered
module 24 · bug 72 · task 64 · decision 19 · relations 453
$ ask any agent: "why was sendfile removed from replication?"
without Byori "I can only see the current implementation. There is no
sendfile code here, so I cannot say whether it ever existed."
with Byori bug:redis-revert-61074b43 (evidence-backed)
Reverted on 2020-06-06: "Implements sendfile for redis."
(commit 9cf500a3f67e)
A revert is the history stating the change was wrong; why it
was wrong is not in the commit, so it is not recorded here.
**On redis's full history: 8/8 questions answered with the graph, 0/8 without it**, where a pass means the answer cited a commit or pull request a reader can open. Ingest took 11 seconds. Reproduce it: [`benchmarks/why.py`](benchmarks/why.py).
Byori preserves that chain rather than a pile of summaries:
incident ──caused_by──> bug ──fixed_by──> decision ──affects──> module
└──supersedes──> previous decision
Vector search recalls a similar paragraph; this recalls the reason — and says what it is standing on. Every memory is labelled `evidence-backed` or `unsourced`, and a decision that something newer replaced comes back marked `stale`, because a memory nobody can check is worth less than one that cites a commit, and last year's decision presented as current is worse than no answer.
What it is made of
byori init |
Builds the sta |
Related Skills
Hearmemory
Shared memory for multi-agent coding: subagents and Codex/Claude Code/Cursor hand-offs share one project memor
Qarinah
Your project remembers across coding agents. Compact, cited project memory for Codex, Claude Code, Cursor, and
Aimux
Run multiple Claude Code, Codex and Gemini subscriptions side by side — shared skills/agents/memory, isolated
Rudder
A TUI for managing coding agents: run Claude Code, Codex and opencode side by side, each in its own isolated j
Team Room
Self-hosted workspace for Claude Code and Codex across your own computers. Shared project memory, files and a
TermCrew
Local multi-agent terminal: run Claude, Codex, Gemini, Cursor Agent and 24 more coding CLIs side by side in re
Related Agents
Part 13: Memory Bridge (Give Coding Agents Your Brain)
When you spawn Codex or Claude Code to build something, they start blind. They don't know your architecture de
Identity Graph Operator
Operates a shared identity graph that multiple AI agents resolve against. Ensures every agent in a multi-agent
Memory Recall
Use this agent PROACTIVELY before starting complex, multi-step implementations or when encountering unfamiliar