Local-first AI agent decision layer

Make your AI agents argue before they ship.

Round Table Workspace gives Codex, Claude Code, and CLI agents a round-table workflow: explore ambiguous ideas, debate risky choices, and run a ship/revise/reject gate before trusting AI-generated work.

Star on GitHub One-minute demo AI feature demo Repo card Use cases Reviewer checklist

$ ./rtw ship-check "Should we merge this AI-generated feature?"

Decision: revise
Panel: product, engineering, risk, user-advocate
Why: useful direction, but public claims and evidence need tightening
Next: run tests, add a visible demo, keep claims local-first unless validated

Who it is forBuilders using Codex, Claude Code, or CLI agents for real product and engineering work.

What it catchesConfident AI output with vague value, missing evidence, risky claims, or unclear next steps.

Why it is differentIt turns one answer into a local review gate with product, engineering, risk, and user perspectives.

/room

Explore first

Turn vague product or engineering questions into structured multi-agent exploration.

/debate

Review risky calls

Escalate mature questions into decision review with explicit tradeoffs and evidence.

ship-check

Gate before shipping

Get a practical ship / revise / reject result before trusting one confident agent.

Try it in 60 seconds

No provider key is required for the default demo path. Star the repo if you want a repeatable review gate before merging AI-generated work.

git clone https://github.com/MarkDonish/round-table-workspace.git
cd round-table-workspace
./rtw demo startup-idea
./rtw ship-check "Should we merge this AI-generated feature?"
./rtw doctor --quick

Workflow

1. AskStart with an idea, change, or launch decision.

2. DiscussUse /room to create a structured exploration trail.

3. DebateUse /debate for product, engineering, risk, and user review.

4. DecideUse ship-check to ship, revise, or reject with boundaries.