idea-lab

Generates and evaluates business and automation ideas through parallel divergence lenses and adversarial critique.

Updated Mar 6, 2026
One-click install
npx skills add https://github.com/ebecerra-developer/ebecerra-web --skill idea-lab-ebecerra-developer
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: idea-lab
Source: https://github.com/ebecerra-developer/ebecerra-web/tree/main/.claude/skills/idea-lab
Command: npx skills add https://github.com/ebecerra-developer/ebecerra-web --skill idea-lab-ebecerra-developer

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Brainstorming sessions for business growth and automation often produce scattered, unvalidated ideas that get forgotten or repeated. This Skill runs a structured divergence-to-convergence pipeline that generates ideas through multiple lenses, filters them with adversarial critique, and persists results in a persistent ledger so discarded ideas never resurface and approved ones improve session over session. ## Core Features & Use Cases - Parallel Divergence with Lenses: Launches multiple ideation subagents in parallel, each applying a distinct lens (productizer, pricing packager, growth, ads, partnerships, contrarian, automation ops) so ideas are diverse by construction. - Adversarial Convergence and Scoring: A viability analyst ranks and kills ideas using a 5-dimension anchored rubric (impact, feasibility, time-to-signal, brand fit, risk) with hard vetoes for legal, ToS, or reputational risk. - Persistent Ledger with Decay: Every idea is tracked in a private ledger with closed states (new, researched, approved, discarded, cold, vetoed), automatic staleness decay, and a veto list that prevents re-proposing rejected concepts. - Use Case: Ask for a quick session on new revenue lines; the Skill checks ledger freshness, runs 4-5 ideation lenses in parallel, lets you veto before any web research is spent, scores survivors, and commits the session report to your private business repo. ## Quick Start Ask the assistant to run an idea-lab session to generate new business or automation ideas for increasing revenue or traffic.

Frequently Asked Questions about idea-lab

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I run a structured business idea brainstorming session with AI?▼

Start an idea-lab session in quick or full mode. The Skill runs a pre-flight check of the ledger and business context, launches parallel ideation agents each with a distinct lens, then presents a deduplicated shortlist for your veto before any research is spent.

How does the idea ledger prevent repeating discarded ideas?▼

Every idea gets a permanent sequential ID and a closed state such as discarded, vetoed, or cold, with a mandatory rejection reason. Before each session a compact veto list is built from these entries so ideation agents never re-propose rejected concepts.

What is the difference between quick and full session modes?▼

Quick mode uses 4-5 lenses, optional research on the top 3, and a single analyst. Full mode uses 6-8 lenses, research on the top 6, a panel of 2-3 analysts with different emphases, refinement plans, and an optional second divergence round.

Can this Skill generate social media content ideas?▼

No. Concrete content angles and pieces for Instagram or Facebook are explicitly out of scope and routed to a separate social scout skill. Idea-lab only keeps the underlying business or distribution mechanism, never the content pieces themselves.

Why does the Skill store session data in a separate private repository?▼

The main web repository is public, so scores, budgets, strategy, and discard reasons are classified as amber data and stored only in a private business repository. Red data like credentials, operational URLs, and third-party PII is never written to any artifact.

What happens when web research fails during idea validation?▼

The idea gets one retry and is then degraded to an unvalidated evidence state rather than blocking the session. Ideas without sources are explicitly marked as hypotheses so unverified claims never contaminate the ledger.