ce-ideate

Generate and evaluate grounded improvement ideas with explicit warrants for software projects.

Updated Apr 26, 2026
One-click install
npx skills add https://github.com/DieStok/ridder_lab_retreat_ai_hackathon_2026 --skill ce-ideate-diestok
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: ce-ideate
Source: https://github.com/DieStok/ridder_lab_retreat_ai_hackathon_2026/tree/main/.agents/skills/compound-engineering/skills/ce-ideate
Command: npx skills add https://github.com/DieStok/ridder_lab_retreat_ai_hackathon_2026 --skill ce-ideate-diestok

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Generate and critically evaluate grounded ideas about a topic within a software project context. This skill helps teams move from vague brainstorming to grounded candidate improvements by requiring explicit warrants and a clear rationale for each idea.

Core Features & Use Cases

  • Generate many ideas grounded in repository context or user-provided constraints.
  • Critically evaluate ideas with explicit warrants (direct:, external:, or reasoned:) and prune to a concise set (5–7 survivors by default).
  • Support surprise-me mode to surface diverse subjects when no subject is specified, while remaining grounded in available context.
  • Produce a structured ideation artifact including titles, summaries, rationale, downsides, confidence, and implementation notes suitable for downstream refinement.
  • Provide a clear quick-start prompt to kick off ideation with a subject or request for surprise.

Quick Start

Provide a subject to ideate on (or choose Surprise me); ce-ideate will generate 5–7 grounded ideas with warrants and a rejection summary.

Frequently Asked Questions about ce-ideate

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

FAQPage Schema
How do I brainstorm grounded software design ideas within my project context?

Grounded software design ideas are generated by analyzing repository context or user-provided constraints to produce candidate improvements. The process requires explicit warrants and clear rationale for each idea, pruning them adversarially to a concise set of survivors.

What is the best way to critically evaluate product improvement ideas for a software project?

Critically evaluating product improvement ideas involves generating multiple candidates with explicit warrants, such as direct, external, or reasoned justifications. Each idea undergoes adversarial critiquing to filter down to 5–7 survivors based on rationale, downsides, and complexity.

How do I structure ideation artifacts to include rationale and implementation notes?

Structured ideation artifacts include titles, summaries, rationale, downsides, confidence, and implementation notes for each candidate. This format provides the explicit warrants and rejection summaries needed to justify why specific ideas survive the filtering process.

Can I generate software project ideas without specifying a subject?

You can generate software project ideas without a specific subject using a surprise-me mode. This mode surfaces diverse subjects grounded in available repository context or user-provided constraints, maintaining the structured warrant and critique evaluation process.

What are the limitations of using adversarial critiquing for software ideation?

A limitation of adversarial critiquing in software ideation is that it strictly filters outputs to 5–7 survivors, rejecting other candidates. This approach requires sufficient repository context or user-provided constraints to generate grounded ideas with explicit warrants.

Does ce-ideate work with user-provided contexts if I don't have a repository?

Yes, ideation works with user-provided contexts if you lack a repository. The process operates on either repo-grounded subjects or user-provided constraints to generate structured candidate ideas with explicit warrants.