research-gatherer

Orchestrates human-gated research workflows for planning, evidence gathering, verification, and synthesis.

Updated Nov 19, 2020
One-click install
npx skills add https://github.com/kwojtasinski-repo/ECommerceApp --skill research-gatherer-kwojtasinski-repo
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: research-gatherer
Source: https://github.com/kwojtasinski-repo/ECommerceApp/tree/main/.github/skills/research-gatherer
Command: npx skills add https://github.com/kwojtasinski-repo/ECommerceApp --skill research-gatherer-kwojtasinski-repo

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve? It prevents unstructured, memory-based answers by enforcing a strict research workflow where every claim is backed by cited evidence from code, docs, config, or external sources, with human approval gates at each stage. ## Core Features & Use Cases - Structured Research Workflow: Classifies requests as internal, external, or mixed, then runs planning, gathering, verification, and synthesis as separate gated phases. - Human Approval Checkpoints: Stops for scope, source, and synthesis approval unless the user explicitly passes --yolo for full autonomy. - Evidence Discipline: Requires file paths and line numbers for code claims, URLs and quotes for web claims, and records rejected-but-relevant findings. - Use Case: Ask whether the ECommerceApp checkout flow matches the payments lifecycle spec; the skill produces a research plan, gathers cited findings into analysis/findings/, cross-checks code against docs, and writes a verified summary. ## Quick Start Ask the research-gatherer to investigate how coupon apply-revert is implemented in ECommerceApp and approve the scope when prompted.

Frequently Asked Questions about research-gatherer

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

FAQPage Schema
How do I run a structured research workflow over a codebase?

Invoke the skill with a research question and it classifies the request, writes a research plan and source manifest, gathers cited evidence into analysis/findings/, verifies cross-source consistency, and synthesizes a summary. Each phase pauses for human approval by default.

What research modes does the research-gatherer support?

It supports full flow, plan-only, sources-only, and synthesize-only modes selected via the --mode flag. Plan-only stops after scope and source strategy; synthesize-only works from existing findings files without new gathering.

Can the research-gatherer run without human approval?

Yes, but only when the user explicitly passes the --yolo flag authorizing full autonomy. Without it, the skill must stop at scope, source, and synthesis checkpoints with an APPROVE / REVISE / ABORT prompt.

When should I not use a research orchestration workflow?

Skip it when you already know the answer, need only a quick lookup or single file read, or want a code change plan or implementation work. The workflow is designed for multi-source evidence-backed research, not quick answers.

How are research findings and citations stored?

Findings are written to files under analysis/findings/, one per source family, with file paths and line numbers for code claims and URLs with quotes for web claims. A summary, verification report, and rejected-findings file complete the output.