Verified:GitHub

research-ops

Orchestrate evidence-first research workflows with labeled facts and dated findings.

Updated Sep 13, 2025

Search Tags:#knowledge management #current events #evidence synthesis #research workflow #lead enrichment #comparative analysis #market recommendations

Ready to run

npx skills add https://github.com/llmh333/employee_management_spring --skill research-ops-llmh333

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: research-ops
Source: https://github.com/llmh333/employee_management_spring/tree/main/.gemini/skills/research-ops
Command: npx skills add https://github.com/llmh333/employee_management_spring --skill research-ops-llmh333

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

The research-ops skill prevents slow, inconsistent, one-off lookups by turning evidence gathering into a repeatable workflow that stays current and clearly separates facts from inference.

Core Features & Use Cases

Evidence-first current-state research flow: Chooses the lightest search lane first, then escalates for synthesis when multiple sources are needed.
Clear evidence boundaries: Labels sourced facts, user-provided context, inference, and recommendations to avoid mixing claims.
Decision-ready outputs: Produces a structured result suitable for comparisons, enrichment, or a recommendation/next move, and flags when a task should become a monitor.

Quick Start

Ask the skill: "Look up and compare the latest options for X, using current public evidence and my notes, then recommend the best next move with clearly labeled sourced facts and dates."

Frequently Asked Questions about research-ops

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

FAQPage Schema

How do I gather current evidence and synthesize it into actionable market recommendations?▼

To gather current evidence for market recommendations, use an evidence-first research workflow that selects the lightest search lane first, escalates for synthesis when multiple sources are needed, and labels sourced facts separately from inference. This ensures decision-ready outputs with freshness-by-date reporting.

What's the best way to enrich company leads with public evidence while avoiding mixed claims?▼

The best way to enrich company leads is by enforcing clear evidence boundaries that label sourced facts, user-provided context, inference, and recommendations separately. This prevents mixing claims and produces structured results suitable for lead intelligence and comparative analysis.

How does evidence synthesis handle recurring research tasks instead of repeated one-off lookups?▼

Evidence synthesis handles recurring research by flagging tasks that should become monitored research instead of repeated one-off lookups. The workflow evaluates whether a question requires ongoing monitoring and recommends setting up a recurring research monitor.

Can I use current events research for comparative analysis with my own notes included?▼

Yes, you can use current events research for comparative analysis by providing user context and notes alongside public sources. The workflow orchestrates evidence gathering across search lanes while enforcing labeled evidence boundaries and freshness-by-date reporting.

When should I escalate from a lightweight search lane to deep research for knowledge management?▼

You should escalate from a lightweight search lane to deep research when multiple sources are needed for evidence synthesis. The workflow starts with the lightest execution first, then escalates to deeper research lanes like market-research or lead-intelligence when comparative analysis requires more comprehensive data.

Does evidence-first research workflow work without external dependencies for current-state findings?▼

Yes, the evidence-first research workflow operates without external dependencies by orchestrating across internal search lanes like exa-search, deep-research, and market-research. It produces current-state findings and recommendations from public sources plus user-provided context using lightweight-first execution.