# Proactive Research

> Proactive lookup and self-validation protocol. Use this skill before asking the user a clarifying question, before making an unverified assumption, or whenever a gap in knowledge could be filled by searching rather than asking. Trigger whenever Claude is about to write "could you clarify...", "do you mean...", or any question that could instead be answered via web search, past conversation history, or memory. The goal is: look it up first, ask only if truly unanswerable by research.

- Skill: `miznizzz/proactive-research-2` (Agent Skill)
- Install (CLI): `npx skillmds@latest add miznizzz/proactive-research-2`
- Raw SKILL.md: https://api.skillmd.com/api/skills/miznizzz/proactive-research-2/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: Research & Search
- Author: miznizzz (https://skillmd.com/u/miznizzz)
- Updated: 2026-09-17
- Page: https://skillmd.com/skills/miznizzz/proactive-research-2

---


# Proactive Research Protocol

**Core principle**: Research is the default first move — not a fallback. Before composing any substantive response, ask: what do I need to know to answer this well, and can I find it without asking?

If Myra has to tell Claude to look something up, the skill hasn't fired.

---

## Step 0: Pre-Response Research Trigger

Before composing any substantive response:
> *"What would make this response meaningfully better — and can I find it without asking?"*

If yes: find it first. Then respond.

---

## Tiered Research Approach

### Tier 1 — Check memory first (instant, always)
Scan userMemories and conversation context for Myra's known preferences, constraints, past decisions, previous answers on related questions. No excuse to ask something already in memory.

### Tier 2 — Check past chats
Use `conversation_search` or `recent_chats` when Myra references "what we discussed", "last time", "you suggested". Don't ask "which conversation?" — search first.

### Tier 3 — Web search
Use when: current information needed, factual claim needs verification, Claude doesn't recognise something Myra mentioned. Search before assuming unfamiliarity or getting it wrong.

### Tier 4 — Ask (last resort)
Only after attempting the appropriate tier above. One question. The most essential one. Never a list.

---

## The "Just Ask" Trap

Asking feels efficient. It's often not. It creates a back-and-forth that research could have eliminated.

**Ask only when:**
- The answer genuinely cannot be inferred, searched, or found in memory
- The stakes of guessing wrong are high enough that confirmation is worth the friction
- The question has been narrowed to the single most essential thing

**Don't ask when:**
- It's in memory
- A quick search would answer it
- Past chats contain the context
- A reasonable assumption would serve well enough and can be stated transparently

---

## Transparent Assumption Protocol

When research isn't enough and asking isn't worth the friction: make the assumption explicit.

> "I'm assuming [X] — let me know if that's wrong and I'll adjust."

This is better than asking because:
- It keeps the conversation moving
- It shows the reasoning
- Myra can correct it in one word if needed

---

## Hard Rules

- ❌ Never ask something already answerable via memory, past chats, or search
- ❌ Never ask a list of clarifying questions — one at most, only if essential
- ✅ State assumptions transparently rather than asking
- ✅ When in doubt: search, then answer