Forge — Engineering & Architecture Deliberation Room
Convenes a panel of Feynman, Ada, Torvalds, Popper, Occam, Nietzsche and Wittgenstein to run an evidence-based, structured debate on architecture, debugging, refactoring and code review.
Dev & CodingIntermediate★ 167⑂ 24AI score 9/10Last updated: Apr 9, 2026
What it does
- Opens the "Forge" deliberation room via
/forgeand auto-assembles a 3–7 persona panel: Feynman (first-principles debugging), Ada (formal systems/abstraction), Torvalds (pragmatic ship-it), Popper (falsification/red-team), Occam (complexity audit), Nietzsche (creative destruction), Wittgenstein (naming/API language precision). - Mandates code evidence before any debate: read relevant files, Grep for key constructs, Glob the module structure,
git loghistory, run tests, inspect dependency manifests — all compiled into a "Forge Evidence Brief". - Runs an 8-step protocol: parse mode → gather evidence → restate the problem plus three probing questions (time pressure, your own leaning, who maintains the code) → Round 1 independent analysis → adaptive depth gate → Round 2 Hegelian cross-examination → synthesis → verdict.
- Produces a Markdown verdict with the architecture recommendation, accepted/rejected trade-offs, a 3-phase implementation path, a risk table, a technical debt ledger (created/paid/deferred), the strongest dissenting position, and a confidence rating.
- Flags:
--quick,--duo,--triad architecture|debugging|refactoring|code-review|api-design|naming|abstraction,--members,--full,--depth auto|full.
Who it's for
- Developers and tech leads facing hard-to-reverse decisions (splitting a monolith, monorepo vs polyrepo)
- Solo devs or small teams who want a deliberate devil's advocate before committing
- Senior engineers who must justify a large refactor or full rewrite with evidence
- Teams short on reviewers who want a red-team lens on PRs and API designs
Examples
/forge --triad architecture "Should we split our 50k-line monolith into microservices?"→ Ada + Occam + Feynman inspect the codebase, deliberate, and return a verdict with a phased implementation path./forge --quick "Is our N+1 query in the user/posts relation worth fixing now?"→ auto-selects the debugging triad and returns a fast verdict plus one concrete next step./forge --duo "Incremental refactor or full rewrite?"→ Torvalds vs Nietzsche dialectic that exposes the irreducible engineering trade-off.
· · · Install guide · · ·
Try it now, no install
Paste this into Claude to use the skill without installing anything.
Read the instructions in this file and follow them to help me: https://raw.githubusercontent.com/geekjourneyx/agora/HEAD/rooms/forge/SKILL.md What I want: (describe your task here)
If Claude can't open the link, open it yourself and paste the contents instead.
↓ If it works for you, download the ZIP below and install it. Then it runs on its own — no pasting each time.
Install in the Claude app (no terminal)
- Download the ZIP with the button below.
- In Claude, open Settings → Capabilities and turn on 'Code execution and file creation'. (one time)
- Go to Customize → Skills → + → 'Upload a skill' and upload the ZIP.
Install in Claude Code
Let Claude do it — paste this into Claude Code
Install the skill I found on Claude Skill Mart. Copy the rooms/forge folder from the GitHub repo geekjourneyx/agora into my ~/.claude/skills/forge/. When it's done, tell me in one line what this skill can do.
Install with a command instead
git clone https://github.com/geekjourneyx/agora.git ~/.claude/skills/agora⚠ This is a third-party skill. Check the source repository before installing.
- Open a terminal.
- Create the skills folder if needed:
mkdir -p ~/.claude/skills - Clone the whole repository:
git clone https://github.com/geekjourneyx/agora.git ~/.claude/skills/agora- The skill lives at
rooms/forge/SKILL.mdbut relies onprotocol/deliberation.mdand thecouncil-*/agora-*sub-agents, so keep the full repo instead of copying a single file.
- The skill lives at
- Verify:
ls ~/.claude/skills/agora/rooms/forge/SKILL.md - Restart Claude Code, then call it:
/forge "How should we split our auth module?". If the command isn't recognized, just ask in plain language: "Use the agora forge skill to…". - Launch Claude Code inside the project you want analyzed so the evidence-gathering step (Read/Grep/git log/tests) actually works.
- The prompt mixes Chinese and English, so add "answer in English" if you want consistent output language.
View source on GitHub ↗License: MIT