research

Produce structured research findings with source citations and confidence levels.

Updated Aug 27, 2026
One-click install
npx skills add https://github.com/mferris77/SpringBoard --skill research-mferris77
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: research
Source: https://github.com/mferris77/SpringBoard/tree/main/vscode-citadel-harness/skills/research
Command: npx skills add https://github.com/mferris77/SpringBoard --skill research-mferris77

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Converts open-ended technical or project questions into concise, evidence-backed findings so teams can make informed next-step decisions without having to do the initial investigative work.

Core Features & Use Cases

  • Structured Findings: Produces discrete findings with What, Source, Relevance, Confidence, and Action fields.
  • Focused Search & Read: Formulates 2–4 targeted queries, reads full pages from credible sources, and stops at 3–6 sources to remain focused.
  • Local Fallbacks: Falls back to local codebase search when web results are irrelevant and writes a persistent research artifact for traceability.
  • Use Cases: Dependency upgrade evaluation, API behavior checks, community best-practices discovery, and codebase-specific investigations.

Quick Start

Ask the research skill to investigate whether upgrading lodash from 4.17.21 to 4.x is safe and produce a findings document with sources and confidence ratings.

Frequently Asked Questions about research

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

FAQPage Schema
How do I evaluate dependency upgrades with source citations?

To evaluate dependency upgrades, this research skill uses 2–4 targeted web queries to gather evidence and produces structured findings with explicit source citations and per-finding confidence levels for informed decisions.

What's the best way to investigate API behavior and codebase-specific questions?

Investigating API behavior is handled by formulating targeted queries and reading full pages from credible sources, falling back to local codebase search when web results are irrelevant to produce structured findings.

How does confidence rating work for technical research findings?

Confidence ratings for technical research are assigned per finding alongside What, Source, Relevance, and Action fields, ensuring each conclusion is explicitly backed by a credible source with a measured confidence level.

Can I generate persistent research artifacts for traceability?

Yes, persistent research artifacts are generated for traceability by writing structured findings to .planning/research/{topic-slug}.md, ensuring the output remains accessible for future reference and project-level decision-making.

Does the research process stop after a certain number of sources?

The research process stops at 3–6 credible sources to remain focused, ensuring the investigation produces concise, evidence-backed findings without excessive information gathering.

Are local codebase searches supported when web research is insufficient?

Local codebase searches are supported as a fallback when web results are irrelevant, allowing the investigation to proceed by reading local files and generating findings for project-level questions.