codegrid
Skill17 starsChanged 5 months agoMITView on GitHub
CodeGrid is a native macOS canvas where multiple coding agents (Claude, Codex, Gemini, Cursor, Grok, shells) run side by side in panes and collaborate via a local agent bus — no tmux, no cloud, no account, no stored API keys. Install this skill when an agent should know how to operate inside a CodeGrid pane, drive the workspace from outside (control socket or codegrid:// deep links), spawn or message sibling agents, or coordinate multi-agent work (delegate, review, pipeline, parallel fan-out, monitor, debate). The differentiator: multiple coding agents collaborating on one canvas, addressable by stable session_id, with a read → message → read protocol built for orchestration.
2 other projects use a version of this.
Adopted by
Most recently taken up by
- BankrBot/openclaw-skillsword for word, first seen Oct 8, 2026
- BankrBot/skillsword for word, first seen Apr 28, 2026
Other versions2 projects use a version of this.
Word for word · 2
- BankrBot/openclaw-skills★ 1.2k100% same
- BankrBot/skills★ 1.2k100% same
Edited · 0
None found.
Wrote their own version · 0
None found.
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
In this repository84 other files it gives its agents.
- AGENTS.mdAGENTS.md
- skills/AGENTS.mdAGENTS.md
- .cursor/rules/codegraph.mdcCursor rule
- 0xworkSkill
- aeon-autoresearchSkill
- aeon-deal-flowSkill
- aeon-deep-researchSkill
- aeon-defi-monitorSkill
- aeon-defi-overviewSkill
- aeon-distribute-tokensSkill
- aeon-hacker-news-digestSkill
- aeon-huggingface-trendingSkill
- aeon-last30Skill
- aeon-monitor-kalshiSkill
- aeon-monitor-polymarketSkill
- aeon-monitor-runnersSkill
- aeon-narrative-trackerSkill
- aeon-on-chain-monitorSkill
- aeon-paper-pickSkill
- aeon-reg-monitorSkill
- aeon-rss-digestSkill
- aeon-skill-evalsSkill
- aeon-skill-repairSkill
- aeon-skill-security-scanSkill
- aeon-token-moversSkill
- aeon-token-pickSkill
- aeon-unlock-monitorSkill
- aeon-vuln-scannerSkill
- agenticbetsSkill
- agent-wormholeSkill
- alchemySkill
- bankr-shopifySkill
- bankrSkill
- token-scam-analysisSkill
- twitter-agentSkill
- baseSkill
- blueagent-x402Skill
- botchanSkill
- botcoin-minerSkill
- capacitrSkill
- cattownSkill
- checkrSkill
- clankerSkill
- darksol-random-oracleSkill
- endaomentSkill
- ens-primary-nameSkill
- erc-8004Skill
- gem-minerSkill
- get-api-docsSkill
- gitlawbSkill
- helixaSkill
- hydrexSkill
- litcoin-minerSkill
- megapotSkill
- moltycashSkill
- nexusSkill
- neynarSkill
- nookplotSkill
- onchainkitSkill
- opensea-apiSkill
- The first 60 shown.
The file
Files
CodeGrid — agent operating manual
CodeGrid runs many coding agents side by side on an infinite 2D canvas. Each
pane is one process (claude, codex, gemini, cursor-agent, grok, or a
shell) with its own working directory, git branch, and stable session_id. The
agent bus lets any pane discover, read, and message any other pane locally.
This skill is the umbrella entry point. Two references cover the full surface:
- Operating CodeGrid — the mental model
(canvas / pane / workspace / agent / bus), MCP tools (
list_agents,read_pane,message_agent), the control-socket JSON-RPC API (agent_list,agent_read,agent_send,open_folder,new_session,new_workspace), thecodegrid://deep-link scheme, and an operating playbook with recipes. - Agent-bus collaboration — the
read → message → readprotocol, identifying agents by role andsession_id, orchestration patterns (delegate, review, pipeline, parallel fan-out, monitor, debate), etiquette and scope safety, loop/runaway prevention, failure recovery, and worked end-to-end examples.
When to load which reference
| You're about to… | Load |
| Discover what's running, open a folder, spawn a session, or drive CodeGrid from outside | using-codegrid |
| Hand work to / consult / review with another agent already running in CodeGrid | codegrid-agent-bus |
| Both | Load using-codegrid first, then codegrid-agent-bus |
Quickstart (inside a CodeGrid pane)
list_agents() # discover panes (session_id, role, status)
read_pane(<session_id>) # see what they're doing — always safe
message_agent(<session_id>, "[from <you>] <self-contained request>")
… wait, then read_pane(<session_id>) for the reply.Quickstart (from outside CodeGrid, e.g. another tool)
SOCK="$(cat ~/.codegrid/socket-path 2>/dev/null || echo ~/.codegrid/socket)"
printf '%s\n' '{"jsonrpc":"2.0","id":1,"method":"agent_list"}' | nc -U "$SOCK"
# Or deep-link the installed app:
open "codegrid://open?path=/abs/path/to/repo&type=codex"
open "codegrid://new"Install
install the codegrid skill from https://github.com/BankrBot/skills/tree/main/codegridSource
- Canonical source of truth lives in the CodeGrid repo under
skills/(skills/using-codegrid/SKILL.md,skills/codegrid-agent-bus/SKILL.md). This Bankr Skills provider is a published copy kept in sync. - App download + full docs: https://codegrid.app.
- Token (Base):
0x6B456E66524aEC1792013eF9DFE87e3F84311ba3— see https://codegrid.app/token.
Safety defaults an agent should respect
- Read before you write.
read_paneis free; never message a busy agent. - Address by
session_id, never pane number — pane numbers shift. - Be explicit about scope. Other agents may be in YOLO/autonomous mode and will act on whatever you send. Say "propose only, do not edit files" when you want analysis.
- One message, then wait. Multiple messages interleave into the target's input box and corrupt it.
- Bound debates and don't spawn unbounded helpers. Converge and report to the user.
- Stay local. Everything is same-machine IPC; there is no remote bus.
Full guidance: see the two reference files.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

