Mmcp.market

research-ideation skill

by pedrohcgs·pedrohcgs/claude-code-my-workflow·1.6k stars·MIT

Generate structured research questions, testable hypotheses, and candidate empirical strategies from a topic, phenomenon, or dataset description. Use when user says "give me research ideas on X", "brainstorm questions about Y", "what could I study with this data?", "I'm looking for a paper idea on...", "generate hypotheses for...". One-shot generation, not multi-turn. For idea-refinement use `/interview-me`.

A100/100content scan

Is the research-ideation skill safe?

Clean: nothing in its files matched our rules. We read 1 file in the folder on 2026-09-28.

No findings.

Install the research-ideation skill

A skill is a folder. Copy it into your agent's skills folder and the agent loads it when the task matches its description. 

git clone --depth 1 https://github.com/pedrohcgs/claude-code-my-workflow.git /tmp/claude-code-my-workflow
mkdir -p ~/.claude/skills
cp -r /tmp/claude-code-my-workflow/.claude/skills/research-ideation ~/.claude/skills/research-ideation
available in every project

In the Claude apps, zip the folder and upload it from the Skills settings. The folder on GitHub

The instructions your agent would load

SKILL.md as published, without the frontmatter. Read it on GitHub

Research Ideation

Generate structured research questions, testable hypotheses, and empirical strategies from a topic, phenomenon, or dataset.

Input: $ARGUMENTS — a topic (e.g., "minimum wage effects on employment"), a phenomenon (e.g., "why do firms cluster geographically?"), or a dataset description (e.g., "panel of US counties with pollution and health outcomes, 2000-2020").

Steps

  1. Understand the input. Read $ARGUMENTS and any referenced files. Check mastersupportingdocs/ for related papers. Check .claude/rules/ for domain conventions.
  1. Generate 3-5 research questions ordered from descriptive to causal:
  • Descriptive: What are the patterns? (e.g., "How has X evolved over time?")
  • Correlational: What factors are associated? (e.g., "Is X correlated with Y after controlling for Z?")
  • Causal: What is the effect? (e.g., "What is the causal effect of X on Y?")
  • Mechanism: Why does the effect exist? (e.g., "Through what channel does X affect Y?")
  • Policy: What are the implications? (e.g., "Would policy X improve outcome Y?")
  1. Tag each RQ with a likely paper type (drawn from methods-referee.md):
  • reduced-form (DiD, IV, RD, event study, synthetic control)
  • structural (estimation of a fully-specified model)
  • theory+empirics (formal model + empirical test of its predictions)
  • descriptive (measurement, data construction, pattern documentation)
  • formal-theory (pure theory, no empirical test in this paper)
  • survey-experiment (vignette, conjoint, list-experiment)
  • unsure (when multiple types are plausible — the user can pick later via /interview-me)

Use .claude/references/discipline-cards.md to bias the distribution by field (econ vs poli-sci default frequencies differ — e.g., poli-sci skews more toward survey-experiment and formal-theory than econ does).

  1. For each research question, develop:
  • Hypothesis: A testable prediction with expected sign/magnitude
  • Identification strategy: How to establish causality (DiD, IV, RDD, synthetic control, etc.)
  • Data requirements: What data would be needed? Is it available?
  • Key assumptions: What must hold for the strategy to be valid?
  • Potential pitfalls: Common threats to identification
  • Related literature: 2-3 papers using similar approaches
  1. Rank the questions by feasibility and contribution.
  1. Save the output to qualityreports/researchideation[sanitizedtopic].md

Output Format

# Research Ideation: [Topic]

**Date:** [YYYY-MM-DD]
**Input:** [Original input]

## Overview

[1-2 paragraphs situating the topic and why it matters]

## Research Questions

### RQ1: [Question] (Feasibility: High/Medium/Low)

**Type:** Descriptive / Correlational / Causal / Mechanism / Policy
**Paper type:** reduced-form / structural / theory+empirics / descriptive / formal-theory / survey-experiment / unsure

**Hypothesis:** [Testable prediction]

**Identification Strategy:**
- **Method:** [the identification approach you would defend in a seminar]
- **Treatment:** [What varies and when]
- **Control group:** [Comparison units]
- **Key assumption:** [the assumption the method's validity rests on, stated so it can be attacked]

**Data Requirements:**
- [Dataset 1 — what it provides]
- [Dataset 2 — what it provides]

**Potential Pitfalls:**
1. [Threat 1 and possible mitigation]
2. [Threat 2 and possible mitigation]

**Related Work:** [Author (Year)], [Author (Year)]

---

[Repeat for RQ2-RQ5]

## Ranking

| RQ | Feasibility | Contribution | Priority |
|----|-------------|-------------|----------|
| 1  | High        | Medium      | ...      |
| 2  | Medium      | High        | ...

Post-Flight Verification (mandatory, CoVe)

Before returning the ideation report, run the Post-Flight Verification protocol from .claude/rules/post-flight-verification.md. Research ideation is hallucination-prone in three specific ways:

  1. Negative-literature claims — "no prior work studies X" is frequently wrong.
  2. Dataset structure claims — "The CPS contains field educ_attain" can be confidently wrong about variable names, coverage years, or restricted-access status.
  3. Estimator feasibility claims — "this works with panel fixed effects" can misstate an identification assumption.

Steps

  1. Extract claims from the draft ideation report: each negative-literature claim, each named dataset with attributed fields, each claimed identification strategy + required data structure.
  2. Generate verification questions per claim. Example: "Has Card & Krueger, Autor, or anyone in the last 10 years studied X? Search Google Scholar + NBER working papers." / "Does IPUMS-CPS include the educ_attain variable 1990–2024?"
  3. Spawn claim-verifier via the Agent tool with subagent_type=claim-verifier, in a fresh context — a named Agent call, not a conversation fork, which would inherit the draft. Hand it claims + questions + source pointers (WebSearch allowed, NBER/SSRN URLs preferred, dataset codebooks preferred). Do NOT include the draft.
  4. Reconcile: PASS → attach green block; PARTIAL → mark uncertain RQs with flags; FAIL → rewrite the affected RQ/hypothesis/strategy.

Skip conditions

  • --no-verify flag
  • User explicitly says "I'll verify the literature myself"

Principles

  • Be creative but grounded. Push beyond obvious questions, but every suggestion must be empirically feasible.
  • Think like a referee. For each causal question, immediately identify the identification challenge.
  • Consider data availability. A brilliant question with no available data is not actionable.
  • Suggest specific datasets where possible (FRED, Census, PSID, administrative data, etc.).

More skills from pedrohcgs/claude-code-my-workflow

  • Aadjudicate-reviewTurn an incoming set of findings — from an AI reviewer, a referee report, a code review, a linter, or a second model — into verified fixes, without letting a confident misread damage correct work. Every finding is a CANDIDATE until checked against the actual source. Use whenever you receive review comments, audit findings, or a critique you did not write yourself, especially when the reviewer is a model or when the volume is too large to check by feel.
  • Aaudit-reproducibilityEnforce the replication-protocol.md rule by cross-checking numeric claims in a manuscript against the actual R / Stata / Python outputs. Report PASS/FAIL per claim against tolerance thresholds. Use before submission and before releasing a replication package.
  • Ablast-radiusBefore and after changing anything shared — a function's return value, a signature, a schema, a label set, a config default, a constant, a file format — find every consumer and actually run them. Catches the change that looks purely additive but silently breaks a contract in a file you never opened. Use when editing shared code, adding a field/column/return element, renaming, changing units or defaults, or touching a pipeline that produces reported numbers.
  • Acapture-environmentSnapshot the computational environment for a replication package — detects the analysis stack (R / Stata / Python) and emits the right lockfiles (renv.lock + sessionInfo.txt, requirements.txt / environment.yml / uv.lock, Stata version + ado package list), records seeds and RNG kind, optionally writes a pinning Dockerfile, and produces a paste-ready "Computational requirements" block. Use when user says "capture the environment", "snapshot my dependencies", "pin the versions", "make a renv.lock / requirements.txt", "make this byte-reproducible", or before releasing a replication package to openICPSR / the AEA Data Editor.
  • AchallengeStress-test a finding against the choices you did not make. Enumerates the discrete forks a competent analyst could have taken (measure definition, sample filter, control set, clustering level, weighting, functional form), runs the specification grid, and reports the distribution rather than a point estimate — then attacks the identifying assumption with named, computable sensitivity statistics. Use when the user says "is this robust", "challenge this result", "specification curve", "multiverse", "how sensitive is this", "what if I'd used a different measure", "stress-test my estimate", or before a result becomes a headline claim. NOT a reviewer of prose or code — it challenges the CLAIM.
  • AcheckpointSave a structured state snapshot before stopping or handing off. Captures the active plan, recent decisions, file pointers (with line numbers), open questions, and the next 1–3 actions into a checkpoint file under `quality_reports/checkpoints/`. Optionally proposes `[LEARN]` entries to add to MEMORY.md. Use when user says "checkpoint", "save state", "snapshot before I stop", "where am I", "wrap up the session for handoff", or before a long break / model switch / collaborator handoff. Companion to (NOT replacement for) the narrative session-log workflow.
  • Acoauthor-briefGenerate a co-author / collaborator handoff brief for a multi-author, multi-machine project — summarizing what changed since the last brief (git delta), the current state of each artifact (manuscript, analysis, slides), open questions, how to reproduce locally, and any restricted-data access steps. Use when user says "coauthor brief", "handoff brief", "bring my coauthor up to speed", "what changed since last week", "onboard a collaborator", "write a handoff for [name]", or before sending a co-author the repo. NOT a commit or a checkpoint — it is the cross-machine, cross-person summary `meta-governance.md` only partially covers.
  • AcommitCommit the current work — runs the quality, consistency and passport gates, branches off main if needed, stages specific files, and writes a commit whose subject states what is now true. Pushes and opens a pull request only with --pr or when the user asks; never merges — a merge happens only when the user explicitly says to merge. Use ONLY on explicit commit intent — user says "commit", "let's commit this", "open a PR", or prefixes with `/commit`. Do NOT auto-invoke on vague end-of-task phrases ("we're done", "wrap up") — those require explicit confirmation first. Never force-pushes or skips hooks.
  • Acompile-latexCompile a Beamer LaTeX slide deck with XeLaTeX (3 passes + bibtex). Use when user says "compile", "build the slides", "rebuild the PDF", "run latex", "render the tex", or asks why a `.tex` file isn't producing a PDF. Operates on `Slides/*.tex`.
  • Acompress-sessionDistill the current conversation into a structured note (decisions made, open questions, file pointers with line numbers, next 1–3 actions) and save to `quality_reports/session_logs/` before auto-compression. Differs from `/checkpoint` (explicit stop-point snapshot) and from auto-compaction (which truncates rather than distills). Use when context is approaching auto-compact threshold, when a long pipeline has accumulated many decisions, or when the user says "compress", "distil this session", "before we hit auto-compact", "structured handoff before context resets".
  • Acontext-statusShow current context status and session health. Use to check how much context has been used, whether auto-compact is approaching, and what state will be preserved.
  • Acreate-lectureCreate a new Beamer lecture `.tex` from source papers and materials, with notation consistency checks and the project's preamble wired in. Use when user says "create a lecture on X", "new lecture from these papers", "start a deck on topic Y", "scaffold a new Beamer file", "build me a lecture from these PDFs". Scaffolds the full deck — NOT for compiling existing `.tex` (use `/compile-latex`).

All agent skills → · MCP servers