superforge

Solid

Concierge and router for the whole making process, from idea to shipped product. Reads intent, hands the work to the right superforge-* skill (brain, biz, brand, ui, dev, test, debug, roast, verify, handoff), and assigns a model tier per subtask across Claude 5, Gemini 3.6, Codex, and Kimi before any agent is dispatched. Use at the start of any build, or when the request spans several of those areas. Use when the user says "let's build", "I want to make", "help me ship", "where do I start", "何か作りたい", "作って", "どこから始める", "一気に進めたい", or runs /superforge.

AI & Automation 1 stars 0 forks Updated today MIT

Install

View on GitHub

Quality Score: 80/100

Stars 20%
10
Recency 20%
100
Frontmatter 20%
70
Documentation 15%
100
Issue Health 10%
80
License 10%
100
Description 5%
100

Skill Content

# Superforge — Concierge & Orchestrator Routes the work, assigns the model tier, keeps the artifacts. The specialist skills own their craft; this skill owns the order of operations and the fact that nothing gets lost between them. --- ## 1. Model & effort tiering — do this before dispatching anything > **判断は Opus 5, 量は Sonnet 5, 雑務は Haiku 4.5, 持久戦は Fable 5** | Tier | Work | Model | |---|---|---| | **A — judgment** | architecture, plan review, root-causing, security review, verifying another agent's claim | Opus 5 | | **A — endurance** | unattended multi-step runs, overnight builds, 10+ step refactors (declare library versions explicitly) | Fable 5 | | **B — volume** | feature implementation, UI building, QA sweeps | Sonnet 5 | | **C — routine** | rote tests, formatting, renaming, log updates | Haiku 4.5 | | **D — bulk text, no repo access** | N variations, summarising pasted text, translation | local `gemini` CLI: `gemini -p "..." -m "gemini-3.6-flash <effort>"` | Never leave every dispatched agent on the session default. That waste is the reason this suite exists. ## 2. Intake For work that opens a new product or feature area, run intake first and write `docs/brief.md` → **`references/intake.md`**. Draft the brief with your own best guesses filled in, then ask the user to correct it. Cap open questions at three per round; assume the rest and log the assumption. Skip intake entirely for bounded tasks inside existing work. ## 3. Route | The user's state | Route to |...

Details

Author
takaoumehara
Repository
takaoumehara/superforge-skill
Created
3 days ago
Last Updated
today
Language
Shell
License
MIT

Similar Skills

Semantically similar based on skill content — not just same category

AI & Automation Solid

superforge-dev

Build multi-component features by decomposing the work and dispatching subagents, assigning a model tier per subtask across Claude 5, Gemini 3.6, Codex, and Kimi. Proposes Subagents versus Agent Teams topology by task complexity and token cost. Use when the user says "implement", "build this feature", "execute the plan", "in parallel", "dispatch agents", "subagents", "which model should", "実装して", "作って", "並列で", "サブエージェント", "プランを実行", "どのモデルで", or runs /superforge-dev.

1 Updated today
takaoumehara
AI & Automation Listed

forge

Designs Claude Code harness artifacts. One triage ladder picks the right surface (skill, hook, path-scoped rule, CLAUDE.md or AGENTS.md entry, workflow script, subagent, MCP); the stance holds the bar every artifact must clear (a gradeable objective, a gap between the model's default and what the owner wants, and a nameable attention shift; the gap is a new idea or source worth operationalizing, a miscommunication to prevent, or an observed failure); per-surface mechanics live in references. Use for creating, reviewing, or refactoring any of these, for distilling articles or docs into one, for skills that under- or over-trigger, hooks that don't fire or fire too often, bloated CLAUDE.mds, and for deciding which surface a behavior belongs on. Triggers on any work on a skill, hook, rule, CLAUDE.md, AGENTS.md, or workflow orchestration script, including debugging triggering or firing semantics and packaging a /command or plugin.

3 Updated today
claudialnathan
AI & Automation Listed

forge

FORGE protocol — world-class multi-model agent swarm and prompt optimiser. Fable engineers a rubric-driven plan (true goal + success rubric + constraints + risks + minimal-sufficient MECE decomposition); Opus critiques & repairs it for MECE/minimality/critical-path; sub-tasks route by tier+kind with effort scaling and real tool use; a perspective-diverse voter panel (with a soundness veto) verifies; failures escalate UP a tier and re-verify; Fable synthesizes and grades the final answer against the rubric. Cost-aware routing: haiku → sonnet → opus → fable. Runs automatically on every message via the UserPromptSubmit hook, and can be invoked manually as /forge <prompt>.

0 Updated yesterday
somudatta06