jhr-workflow - SKILL.md Agent Skill

name: jhr-workflow description: Use when deciding which jhr-* sub-skill to invoke next, or when sequencing manuscript work from topic selection through rebuttal for a Journal of Human Resources (JHR) submission. Routes — it does not replace — the specialized skills.

JHR Workflow Router (jhr-workflow)

Overview

This is the router. It does not replace any specialized skill. It tells you which jhr- skill to use at the current stage* of a manuscript aimed at the Journal of Human Resources.

Default assumption: unless told otherwise, treat the target as JHR — a leading empirical-microeconomics journal founded in 1965, published by University of Wisconsin Press and housed at the UW-Madison Institute for Research on Poverty (IRP). It covers labor, development, health, education, discrimination, and retirement economics with a strong applied-micro, causal-identification, and public-policy orientation. Crucially, despite its name, JHR does NOT consider management or personnel ("HR management") research — and the nonrefundable submission fee is not returned for out-of-scope papers. The Editor is Anna Aizer (Brown) as of 2026-06-01; Michael Lovenheim (Cornell) becomes Editor on 2026-07-01 — re-verify on the live editorial-board page.

When to trigger

The user asks "what should I do next?"
A draft is handed over and the current bottleneck needs diagnosing
Work is ping-ponging between empirics, writing, and the response letter
A JHR decision letter (desk reject or R&R) just arrived

Routing table

Current symptom	Next skill
Unsure the paper is even in scope (or it smells like "HR management")	`jhr-topic-selection`
Contribution relative to prior policy-relevant work is fuzzy	`jhr-literature-positioning`
Causal claim rests on OLS + controls	`jhr-identification-strategy`
Estimates exist but specs/sample/reconciliation are unsettled	`jhr-data-analysis`
The "why JHR / why now" framing is weak	`jhr-contribution-framing`
Tables are dense; exhibits not policy-legible	`jhr-tables-figures`
Prose buries the policy lesson	`jhr-writing-style`
Accepted-stage CC0 deposit, archive plan, DAS	`jhr-replication-and-data-policy`
Want to understand desk-reject / single-anonymous review	`jhr-review-process`
Ready to submit via msubmit.net; need a preflight	`jhr-submission`
Received an R&R; need a response-letter strategy	`jhr-rebuttal`

Default order

jhr-topic-selection 2. jhr-literature-positioning 3. jhr-identification-strategy 4. jhr-data-analysis 5. jhr-contribution-framing 6. jhr-tables-figures 7. jhr-writing-style 8. jhr-replication-and-data-policy 9. jhr-review-process 10. jhr-submission 11. jhr-rebuttal

jhr-writing-style is a late polish — do not rewrite the intro before identification and reconciliation are settled.

Diagnosis questions before routing

Ask these in order; the first "no" usually names the bottleneck skill:

Is the question empirical microeconomics in a JHR field, with a policy lever — and definitely not HR-management? (jhr-topic-selection)
Can the draft name the closest prior estimate and why magnitudes differ? (jhr-literature-positioning)
Does a referee-grade design exist: rollout timing, cutoff, lottery, or instrument, with the matching diagnostics planned? (jhr-identification-strategy)
Are estimates produced with venue-default estimators, assignment-level clustering, and a reconciliation table? (jhr-data-analysis)
Are the archive footnote, disclosure statements, and 40-page budget ready? (jhr-replication-and-data-policy, then jhr-submission)

Decision-letter triage

Desk release (no referees): diagnose scope vs. framing — re-run jhr-topic-selection; if scope was fine, the desk screen failed on legibility, so route to jhr-contribution-framing before any new venue.
R&R: go straight to jhr-rebuttal, but expect loops back into jhr-data-analysis for comparative estimation and into jhr-tables-figures for new diagnostics.
Reject after review: harvest the reports; they may be attachable at the next venue, and JHR's review-recycling logic works in reverse too.

Worked routing pass

Incoming draft: staggered state paid-leave rollout, TWFE only, no event study, no archive plan, intro leads with three pages of literature. Routing emitted:

【Stage】pre-submission, empirics unsettled
【Bottleneck】TWFE with staggered adoption; pre-trends never shown
【Route】jhr-identification-strategy -> jhr-data-analysis (robust estimator,
        event study, reconciliation table)
【Then】jhr-contribution-framing (move design + magnitude to page 1)
【Parallel】jhr-replication-and-data-policy (archive footnote before preflight)
【Hold】jhr-writing-style until estimates stabilize

Emit this block whenever the user asks "what next" — one stage, one bottleneck, one ordered route, holds made explicit.

Cross-skill invariants

Whatever route is taken, three artifacts must exist before jhr-submission runs: the reconciliation table, the design-diagnostic exhibit (event study, first stage, or density test), and the archive-plan footnote. If a routing pass ends without all three on the board, schedule the owning skill explicitly.