Forge

Install in one command

mkdir -p ~/.claude/skills/forge && curl -fsSL https://eliteaiadvantage.com/skills/forge/SKILL.md -o ~/.claude/skills/forge/SKILL.md

Forge is a sister skill to Verdict, built for tool and product ideation. Five contributors take a raw 'what if we built X' thought and either turn it into a buildable MVP scope, reshape it into a sharper version, or kill it. Each contributor commits to a position from a different angle: brand positioning, the buyer's chair, the differentiator hunt, the delivery cost, and where it sits in the funnel. After they argue, Claude synthesizes one verdict, an MVP scope, the distinguishing element a competitor can't copy, and a single concrete next action.

It's built for the moment a half-formed tool concept needs to be pressure-tested before any code gets written. Most tool ideas die one of three deaths: clones anyone could publish, no funnel role, or too ambitious to ship. Forge surfaces the death early so it doesn't waste a build cycle.

• →Forces a build / reshape / kill verdict instead of an open-ended brainstorm that produces nothing shippable.
• →Names the distinguishing element, the line a competitor cannot copy, before any code gets written.
• →Outputs an MVP scope tight enough to brief a developer with tomorrow.
• →Maps the tool to a funnel role and names what gets captured (email, intent signal, scoping click).
• →Flags the strongest dissent so you know what to watch for once the tool is live.
• →Sister to /verdict (decisions) and /prospect (idea generation), same five-voice family, calibrated for build calls.

• •Pressure-testing a half-formed tool or product idea before writing code.
• •Choosing between forging an MVP, reshaping the concept, or killing it.
• •Naming the moat for a tool that risks being a clone of HubSpot or McKinsey output.
• •Sizing an MVP scope you can ship in a week.
• •Mapping a new tool to its funnel role and intent capture.
• •Auditing whether a tool concept is on-brand for your wedge or dilutes it.

Without this skill, Claude tends to validate tool ideas with balanced 'pros and cons' that ship nothing. Forge reshapes the response into five committed contributors plus a synthesis that lands on build, reshape, or kill, with an MVP scope, a moat, a funnel role, and a 48-hour next action. You stop getting feasibility reports and start getting buildable specs or clean kills.