Research Planning

Creates structured research plans with sub-questions, query strategies, claim budgets, and source tier rules.

Updated Jan 29, 2026
One-click install
npx skills add https://github.com/sparkst/sparkry-claude-skills --skill research-planning-sparkst
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: Research Planning
Source: https://github.com/sparkst/sparkry-claude-skills/tree/main/plugins/research-workflow/skills/research/plan
Command: npx skills add https://github.com/sparkst/sparkry-claude-skills --skill research-planning-sparkst

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Research requests often start without structure, leading to unfocused web searches, unverified claims, and no clear definition of done. This Skill enforces a systematic planning phase that decomposes any research topic into sub-questions, targeted queries, claim budgets, and source quality rules before any searching begins. ## Core Features & Use Cases - Structured Plan Generation: Produces a research/plan.json file with topic, objective, 3-7 sub-questions, search queries, tier rules, and success criteria following a defined JSON schema. - Claim Budget & Source Tiers: Limits load-bearing claims (≤10 recommended) and enforces Tier-1 source requirements for high-stakes claims like market sizing. - Query Strategy Library: Includes advanced search operator patterns for Tavily and Brave, canonical domain targeting, noise filtering, and temporal strategies. - Use Case: A founder asks to research the AI coding agents market for positioning. The Skill produces a plan with 5 sub-questions, 10 targeted queries, a claim budget of 10, and Tier-1 requirements for market size data before any web search runs. ## Quick Start Create a research plan for the AI coding agents market with sub-questions, search queries, and a claim budget before starting any web searches.

Frequently Asked Questions about Research Planning

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

FAQPage Schema
How do I structure a research plan before web searching?▼

Decompose the topic into 3-7 specific sub-questions using the 5W1H framework, generate 2-3 targeted search queries per question, set a claim budget of 10 or fewer, and define tier rules for source quality. Save the result as a research/plan.json file for confirmation before execution.

What is a claim budget in research planning?▼

A claim budget limits the number of load-bearing claims in a deliverable, with 10 recommended for most research. Quick research uses 5 claims, while deep research may justify 10-15. This constraint maintains signal over noise in the final output.

What are Tier-1 sources for market research?▼

Tier-1 sources are primary and authoritative: government agencies like census.gov and sec.gov, research firms like Gartner, IDC, and McKinsey, peer-reviewed journals, and official technical documentation. High-stakes claims require at least two independent Tier-1 sources.

How do I filter noise from web search results?▼

Exclude content farms, affiliate review pages, and low-authority domains using operators like -site:medium.com, -inurl:review, and -inurl:best. Target canonical domains with site: filters and use freshness parameters to prioritize recent data.

When should Tier-2 sources be acceptable in research?▼

Tier-2 sources like the Wall Street Journal, Ars Technica, and CB Insights are acceptable for medium-stakes claims such as features, pricing, and sentiment, or for exploratory low-stakes research. They can supplement but not replace Tier-1 sources for high-stakes claims.