Agent Skills
› brycewang-stanford/Awesome-Journal-Skills
› red-workflow
red-workflow
GitHub用于规划《经济动态评论》(RED) 稿件的全生命周期工作流。涵盖从动态量化范围确认、文献定位、模型构建、代码归档到投稿及修改的完整路径，并路由至各专项技能以辅助执行。
Trigger Scenarios
开始撰写针对 RED 期刊的论文
不确定下一步该执行哪个 RED 特定步骤
检查是否遗漏费用或代码归档等关键要求
Install
npx skills add brycewang-stanford/Awesome-Journal-Skills --skill red-workflow -g -y
SKILL.md
Frontmatter
{
"name": "red-workflow",
"description": "Use when mapping the end-to-end lifecycle of a Review of Economic Dynamics (RED) manuscript — from confirming dynamic\/quantitative scope, through the current Elsevier Guide's USD 195 fee and ScienceDirect\/Editorial Manager submission, the single-anonymized two-reviewer process, the code-first data\/code archive plus Option C data statement, to the revise-and-resubmit. Orchestration; it routes to the other red- skills rather than drafting content."
}
RED Manuscript Workflow (red-workflow)
When to trigger
- Starting a paper aimed at the Review of Economic Dynamics and wanting the full path
- Unsure which RED-specific step comes next or which sibling skill to open
- Sanity-checking that nothing RED-specific (fee, code archive, single-anonymized norms) is missed
The lifecycle (RED-specific)
- Scope check — Is the question dynamic and quantitative, studied through a dynamic model
(theoretical, computational, or empirical)? RED's scope is defined by method/lens, not subfield. →
red-topic-selection - Position — Place the paper in the dynamic-economics / SED literature. →
red-literature-positioning - Frame the contribution — What is the marginal quantitative/dynamic advance? →
red-contribution-framing - Identification / model logic — Assumptions and regularity conditions (theory), or causal
design (empirical). →
red-identification-strategy - Analysis — Calibration, moment-matching, estimation, numerical solution discipline. →
red-data-analysis - Exhibits — IRFs, moment tables, model-vs-data figures. →
red-tables-figures - Write — RED house style, author-year references, ≤250-word stand-alone abstract, 1–6 keywords. →
red-writing-style - Stage replication — Build the data/code archive early; RED's policy is code-first. →
red-replication-and-data-policy - Understand review — Desk screen, single-anonymized, ≥2 reviewers when sent out; plan around fast-review norms without assuming a guaranteed decision date. →
red-review-process - Submit — Pay the USD 195 fee (USD 100 all-student), submit via ScienceDirect / Editorial Manager. →
red-submission - Revise — Response-to-referees after an R&R; resubmissions are fee-exempt. →
red-rebuttal
Anti-patterns
- Treating RED like a generalist top-5 venue and ignoring its method-defined scope
- Forgetting the per-submission fee gates the review (no fee paid → no review)
- Leaving the replication archive or Option C data statement to the accepted stage when the policy expects code-first discipline
Router diagnostics
Ask these questions before choosing the next skill:
- Is the dynamic mechanism unclear? Use
red-topic-selectionorred-contribution-framing. - Is the model disciplined by too few moments or free parameters? Use
red-data-analysis. - Are assumptions, existence, uniqueness, or accuracy weak? Use
red-identification-strategy. - Are exhibits static or uninformative about dynamics? Use
red-tables-figures. - Is the archive missing a run-all path, seeds, or runtime? Use
red-replication-and-data-policy.
The right RED route is usually the step that makes the dynamic model more auditable.
Stage-gate tracker
Keep one block per project and tick gates as they clear:
RED PIPELINE — [paper short title]
G1 scope: dynamic mechanism named; model class chosen [date/✓]
G2 positioning: lineage map written; closest model identified [ ]
G3 discipline: calibration table complete; every target sourced [ ]
G4 computation: solver chosen; accuracy check passing [ ]
G5 exhibits: mechanism figure + fit table drafted [ ]
G6 archive: run-all reproduces exhibits; readme.txt drafted [ ]
G7 manuscript: abstract ≤250w; author-year refs; 1–6 keywords [ ]
G8 submission: fee path decided; ScienceDirect upload checked [ ]
G9 review: desk cleared; reviewer-clock expectations logged [ ]
G10 revision: point-by-point reply; archive re-run and re-sent [ ]
Gates G3–G6 are where RED differs most from a generalist target: they exist because SED-community referees re-derive calibrations and sometimes re-execute code.
Sequencing rules specific to RED
- Build the archive (G6) in parallel with the analysis, not after acceptance — the data/code policy bites before final acceptance, and retrofitting seeds and execution order into months-old solver code is the classic schedule killer for simulation-heavy papers.
- Do not draft the introduction before G2 and G3 exist; an intro written without a lineage map and a calibration table tends to over-claim and gets rewritten after the moments come in.
- Budget compute as a project resource: a global-solution model whose counterfactual takes days to run limits how many referee experiments a two-round revision can absorb.
Failure-mode router
| Observed failure | Gate | Skill |
|---|---|---|
| Headline number moves when the grid is refined | G4 | red-identification-strategy |
| Untargeted moments badly missed | G3 | red-data-analysis |
| Reader cannot restate the mechanism after the intro | G1/G2 | red-contribution-framing |
| Archive fails a clean-machine rerun | G6 | red-replication-and-data-policy |
Output format
[Current stage] idea / model / calibration / estimation / writing / submission / review / R&R
[Dynamic bottleneck] scope / mechanism / calibration / computation / archive / response
[Next RED skill] <red-* skill>
[Hard blocker] <fee, archive, model credibility, review risk, or none>
[Next action] <single concrete task>
Supplementary resources
../../resources/official-source-map.md— official RED URLs behind every fact../../resources/external_tools.md— Dynare, Julia/SSJ, FRED and reproducibility tooling
Version History
- 1839142 Current 2026-07-05 14:21
