skills.rest
Dozyboy
DozyboyCOMMUNITY
Explore all skills by this author →
GitHub

hs:discover

Shapes ambiguous problems into discovery briefs via research and brainstorm chains.

Updated Jul 19, 2026
Search Tags:#planning#research#requirements#workflow orchestration#discovery#brainstorm#problem framing
One-click install
npx skills add https://github.com/Dozyboy/VSF --skill hs-discover-dozyboy
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: hs:discover
Source: https://github.com/Dozyboy/VSF/tree/main/Day2_VSF/Demo1/harness/plugins/hs/skills/discover
Command: npx skills add https://github.com/Dozyboy/VSF --skill hs-discover-dozyboy

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve? Vague feature ideas and ambiguous problems often lead teams to plan and build the wrong thing. This Skill turns an unclear problem description into a structured discovery brief with evidence, option analysis, trade-offs, and open questions before any planning begins. ## Core Features & Use Cases - Research and brainstorm chain: Orchestrates hs:research for evidence gathering, then hs:brainstorm diverge/critique/converge passes to generate and evaluate 2-4 approaches. - Structured discovery brief: Writes a standardized brief to plans/<slug>/discovery-brief.md covering problem framing, constraints, option space, chosen direction, risks, and explicit out-of-scope items. - Tiered handoff to planning: Assesses scope and routes the brief to hs:plan with the right flags (--deep/--tdd for complex work, --fast for tiny changes), recommending a /clear for context isolation. - Use Case: A user says "I want to add caching but I'm not sure where." The Skill interviews for constraints, researches prior art, brainstorms options, critiques them, and produces a brief recommending a direction with trade-offs for the planning stage. ## Quick Start Ask the AI to run hs:discover with a description of your problem or feature idea, adding --quick if the problem is simple and well understood.

Frequently Asked Questions about hs:discover

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

FAQPage Schema
How do I turn a vague feature idea into a plan?▼

Run hs:discover with your problem description. It interviews you for constraints, researches evidence, brainstorms 2-4 approaches, critiques them, and writes a discovery brief that hs:plan consumes to produce an implementation plan.

When should I use discovery before planning?▼

Use discovery when the problem is described by outcome rather than solution, when two or more fundamentally different approaches exist, or when technical constraints are unclear. Skip it when the solution is already confirmed or only one reasonable implementation exists.

What does the --quick flag do in hs:discover?▼

The --quick flag skips the research step and replaces the full diverge/critique/converge brainstorm chain with a single fast brainstorm pass. Use it for small, low-risk problems where no open technical questions require external research.

What happens when discovery gets stuck with no viable options?▼

After two full brainstorm rounds with no surviving option, or when the frame still spans two or more unresolved dimensions, the Skill routes to hs:problem-solving for a reframe instead of re-running the same brainstorm loop.

Does hs:discover write code or create the implementation plan?▼

No. Its only output is a discovery brief in markdown under plans/. It does not write code, edit harness files, or create the plan; the brief is handed off to hs:plan with a recommended tier and flags.

Related Skills

Explore More Software Engineering Skills→
rupertsworldrupertsworld
1

code-init

Analyze unfamiliar codebases by reading docs, mapping structure, and summarizing architecture.

Community
Intermediate
dmedina-devdmedina-dev
1

brainstorming-workflow

Orchestrate teammate-driven feature development from discovery through PR creation.

Community
Advanced
openJiuwen-aiopenJiuwen-ai
64

evolve-design

Designs and launches checkpointed evolutionary search runs that iteratively optimize text-shaped candidates against measurable scores.

Official
Advanced
SensLiaoSensLiao

gsd-fast

Execute one-sentence tasks directly in the current context without subagents.

Community
Basic