brycewang-stanford/Auto-Empirical-Research-Skills / skills/15-Felpix-Studios-social-science-research/skills/new-project
new-project
VerifiedStart a new research project by conducting a structured interview to formalize a research idea, then generates research questions with identification strategies and a project spec. Make sure to use this skill whenever the user wants to develop or document a new research idea — not to search for literature or data. Triggers include: "new project", "start research", "I have an idea", "help me develop this", "I want to study X", "help me formalize this idea", "what's my research question", "what identification strategy should I use", "write up my project idea", or when the user describes a topic they want to turn into a paper.
| File | Kind | Size |
|---|---|---|
| SKILL.md | skill md | 6.5 KB |
SKILL.md
View on GitHub---
name: new-project
description: >-
Start a new research project by conducting a structured interview to formalize a research idea, then generates research questions with identification strategies and a project spec. Make sure to use this skill whenever the user wants to develop or document a new research idea — not to search for literature or data. Triggers include: "new project", "start research", "I have an idea", "help me develop this", "I want to study X", "help me formalize this idea", "what's my research question", "what identification strategy should I use", "write up my project idea", or when the user describes a topic they want to turn into a paper.
argument-hint: "[brief topic or 'start fresh']"
allowed-tools: ["Read", "Grep", "Glob", "Write"]
---
# New Research Project
Formalize a research idea into a concrete project specification with testable hypotheses and empirical strategies.
**Input:** `$ARGUMENTS` — a topic, phenomenon, dataset, or "start fresh" for open-ended exploration.
This skill runs in **three phases**. Phase 1 is conversational — ask one or two questions at a time and wait for responses. Phases 2 and 3 run automatically after the interview.
---
## Phase 1: Research Interview
**Goal:** Draw out the researcher's thinking and establish a clear research question.
Ask questions **one or two at a time**. Build on each answer before moving to the next phase. Do NOT use AskUserQuestion — ask directly in your response. A good interview runs 4–6 exchanges.
### Question Bank (select and adapt based on context)
**The Puzzle (start here):**
- "What phenomenon or puzzle are you trying to understand?"
- "What do you observe in the data / world that doesn't fit the standard explanation?"
**Why It Matters:**
- "Why does this matter? Who should care about the answer?"
- "Is there a policy lever here, or is this more about understanding a mechanism?"
**Theoretical Motivation:**
- "What's your intuition for why X happens — what's the mechanism?"
- "What would standard theory predict? Do you expect to find something different, and why?"
**Data and Setting:**
- "Do you have data in mind, or are you open on the data source?"
- "Is there a specific context, time period, country, or institutional setting you're focused on?"
**Identification:**
- "Is there a natural experiment, policy change, or discontinuity you could exploit?"
- "What's the biggest threat to a causal interpretation — what would a skeptic say?"
**Expected Results + Contribution:**
- "What would you expect to find? What would genuinely surprise you?"
- "What existing papers are closest to this? What gap does yours fill?"
### When to Stop Interviewing
Move to Phase 2 when you have:
- A clear research question (one sentence)
- At least one plausible identification strategy
- Some sense of what data exists or is needed
- The motivation / contribution
If after 3 exchanges the user keeps giving vague answers, move to Phase 2 anyway and flag the open questions.
---
## Phase 2: Research Ideation
**Goal:** Generate 3–5 structured research questions covering the full range from descriptive to causal.
Announce the transition: *"Great — I have enough to generate a structured set of research questions. Let me build that out now."*
Then generate **3–5 research questions** ordered by type:
| Type | What It Asks |
|------|-------------|
| **Descriptive** | What are the patterns? How has X evolved? |
| **Correlational** | What factors are associated with X, controlling for Z? |
| **Causal** | What is the causal effect of X on Y? |
| **Mechanism** | Through what channel does X affect Y? |
| **Policy** | Would intervention X improve outcome Y? |
**For each RQ, develop:**
- **Hypothesis** — testable prediction with expected direction/magnitude
- **Identification Strategy:**
- Method (DiD, RDD, IV, synthetic control, etc.)
- Treatment (what varies, when, where)
- Control group (comparison units)
- Key assumption (parallel trends, exclusion restriction, etc.)
- Main robustness checks (pre-trends test, placebo, etc.)
- **Data requirements** — what variables, time period, geography, unit of observation
- **Key pitfalls** — 2 main threats to identification + mitigations
- **Related work** — 2-3 papers using similar approaches (name only, no fabrication)
**Rank the questions** by feasibility × contribution:
| RQ | Feasibility | Contribution | Priority |
|----|-------------|-------------|----------|
| 1 | High | High | ★★★ |
| 2 | High | Medium | ★★ |
| ... | ... | ... | ... |
---
## Phase 3: Save Project Spec
Produce the unified project spec document and save it.
**Save to:** `quality_reports/project_spec_[sanitized_topic].md`
```markdown
# Research Project: [Working Title]
**Date:** [YYYY-MM-DD]
**Researcher:** [from CLAUDE.md if available]
---
## Research Question
[Single clear sentence]
## Motivation
[2–3 paragraphs: why this matters, theoretical context, policy relevance, what the answer would change]
## Research Questions
### RQ1: [Question] — Priority: ★★★ (Feasibility: High / Contribution: High)
**Type:** Causal
**Hypothesis:** [Testable prediction with expected sign]
**Identification Strategy:**
- **Method:** [e.g., Staggered DiD with Sun–Abraham estimator]
- **Treatment:** [What varies and when]
- **Control group:** [Comparison units]
- **Key assumption:** [e.g., Parallel pre-trends conditional on controls]
- **Robustness:** [Pre-trends test, placebo outcomes, alternative control groups]
**Data Requirements:**
- [Dataset or data type needed]
- [Key variables: treatment proxy, outcome, controls]
- [Time period and geography]
**Key Pitfalls:**
1. [Threat + mitigation]
2. [Threat + mitigation]
**Related Work:** [Author (Year)], [Author (Year)]
---
[Repeat for RQ2–RQ5]
---
## Priority Empirical Strategy
[1 paragraph recommending the single highest-priority RQ and why, with the specific identification approach]
## Open Questions
[Issues raised in the interview that need further thought before committing to a strategy]
---
## Suggested Next Steps
1. **`/lit-review [topic]`** — Search the literature for related work and citation chains
2. **`/data-finder [topic]`** — Find and assess datasets for the priority RQ
3. Once data is secured: **`/data-analysis`** to begin analysis
```
---
## After Saving
Tell the user:
- The spec is saved to `quality_reports/project_spec_[topic].md`
- Recommended next step: `/lit-review [topic]` to build the literature foundation
- Then: `/data-finder [topic]` to identify and assess data sources
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| reverse-engineer-anythingmorluto/reaReverse engineer native, managed, Electron/JavaScript, packaged, firmware, and browser targets with REA, and build agent environments (AgentEnv bundles) of web applications from a live instance or a capture, and of Linux desktop apps with a task from a request. Use shipped-artifact or requested runtime evidence to explain features, compare versions, decompile code, or guide a reconstruction. Skip REA for ordinary source-repository architecture analysis. | Stars: 472k | Updated | Claude CodeCodex | Verified | Install |
| diagnosing-superpowersobra/superpowersUse when a superpowers session went wrong and your human partner wants to know why — repeated work, ignored plans, stumbles, poor results, a skill that didn't fire, "it took too long", "why is it so expensive", "what is it doing" — or wants to build a bug report for the superpowers maintainers, for the current session or a past one identified by id or path, on any harness. | Stars: 323k | Updated | Claude CodeCodex | Verified | Install |
| dispatching-parallel-agentsobra/superpowersUse when facing 2+ independent tasks that can be worked on without shared state or sequential dependencies | Stars: 323k | Updated | Claude CodeCodex | Verified | Install |
| executing-plansobra/superpowersUse when executing an implementation plan in the current session as the implementer yourself — your human partner chose inline execution, or no subagent tool is available | Stars: 323k | Updated | Claude CodeCodex | Verified | Install |