next-idea

Generates and files validated product improvement ideas as locked GitHub issues.

5.4k|571|Updated Nov 1, 2025
One-click install
npx skills add https://github.com/breaking-brake/cc-wf-studio --skill next-idea
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: next-idea
Source: https://github.com/breaking-brake/cc-wf-studio/tree/main/.claude/skills/next-idea
Command: npx skills add https://github.com/breaking-brake/cc-wf-studio --skill next-idea

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Maintaining a steady pipeline of high-quality, user-focused improvement proposals for cc-wf-studio requires manual ideation and triage. This Skill automates one unattended ideation iteration, ensuring only verified, valuable ideas enter the backlog without ever touching code.

Core Features & Use Cases

  • Autonomous Ideation Loop: Runs an orient → invent → judge → file cycle that reads the implementation plan, progress log, and open issues before proposing 3-5 fresh improvements.
  • Strict Value Bar: Filters proposals against five criteria (value axis alignment, user-noticeable impact, single-PR shippability, safety, and code-verified premises) so filler never gets filed.
  • Locked Issue Filing: Creates up to 3 GitHub issues labeled idea and locks them immediately so the body remains a trusted spec for the downstream next-task implementation skill.
  • Use Case: A maintainer schedules this Skill to run nightly; it detects that the idea queue has capacity, verifies a gap in the MCP tool flow by reading the code, and files a locked issue describing a user-facing improvement with verified code locations.

Quick Start

Ask the agent to run the next-idea skill to generate and file new improvement ideas for the repository.

Frequently Asked Questions about next-idea

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

FAQPage Schema
How do I automate product idea generation with an AI agent?

Run this Skill to execute one ideation iteration: it reads the project plan and open issues, invents 3-5 proposals, judges them against a value bar, and files winners as GitHub issues. It never writes code, so it pairs with a separate implementation skill.

How does the skill decide which ideas to file as GitHub issues?

Each proposal must pass five checks: alignment with a value axis in IMPLEMENTATION_PLAN.md, a user-noticeable outcome, shippability in one PR, safety with no breaking changes, and verification against actual source code. At most 3 passing ideas are filed per run.

Can this skill implement the ideas it proposes?

No. The Skill is strictly read-only toward the repository and never commits, branches, or opens PRs. Implementation is handled by the separate next-task skill, which consumes the idea queue this Skill fills.

What happens when too many idea issues are already open?

The Skill applies queue back-pressure: if 5 or more issues labeled idea are already open, it files nothing and ends the iteration early. An empty iteration is considered a valid outcome.

Why are the created GitHub issues locked immediately?

Locking restricts comments to collaborators, keeping the issue body authored only by the owner or the automation loop. This protects the spec that the next-task skill builds from while still letting the human maintainer comment or close the issue.