菜单
Utility workflows for compile, bibliography validation, journal/log inspection, runtime smoke tests, memory updates, bootstrap tasks, and infrastructure refresh. Use when the user asks for mechanical project maintenance rather than substantive research work.
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
基于 SOC 职业分类
End-to-end data analysis using Data-engineer, Coder, and coder-critic. Use when the user asks to clean data, build a panel, estimate a design, reproduce a strategy, or perform a dual-language replication.
Discovery phase workflows for research interviews, literature review, data discovery, and ideation. Use when the user asks to scope a topic, review the literature, find datasets, or generate viable research designs.
Full research pipeline from idea to paper for the Codex-native Clo-Author. Use when starting a new empirical research project from scratch or bootstrapping this template into a live project.
All quality reviews for manuscripts, methods, code, talks, and replication packages. Use when the user asks for code review, paper review, causal audit, proofreading, peer review simulation, stress testing, or a weighted quality score.
Design identification strategy or pre-analysis plans using Strategist plus strategist-critic. Use when the user asks for a causal design, robustness plan, falsification tests, estimator choice, or a PAP.
Create, audit, or compile presentations derived from the paper using Storyteller plus storyteller-critic. Use when the user asks for a job-market talk, seminar, short talk, lightning talk, or slide audit.
| name | tools |
| description | Utility workflows for compile, bibliography validation, journal/log inspection, runtime smoke tests, memory updates, bootstrap tasks, and infrastructure refresh. Use when the user asks for mechanical project maintenance rather than substantive research work. |
Run lightweight utility workflows for project maintenance and infrastructure.
AGENTS.mdplugins/clo-author/references/runtime-activation.mdplugins/clo-author/references/codex-workflow.mdplugins/clo-author/references/logging.mdMEMORY.md when the task concerns durable learningsUse the active paper chain:
cd paper && TEXINPUTS=preambles:$TEXINPUTS xelatex -interaction=nonstopmode main.tex
biber main
TEXINPUTS=preambles:$TEXINPUTS xelatex -interaction=nonstopmode main.tex
TEXINPUTS=preambles:$TEXINPUTS xelatex -interaction=nonstopmode main.tex
Cross-check \cite{} keys against Bibliography_base.bib and report:
Use plugins/clo-author/references/runtime-activation.md as the contract and report whether the session is active or blocked before claiming the plugin is available.
If the home-level skill link is missing, install or verify it with:
bash scripts/install_codex_skill_link.sh
Summarise the latest plan, session log, and research journal when the user asks for current state or recovery context.
When updating memory, write concise [LEARN:*] entries grounded in actual mistakes or durable workflow discoveries.
Refresh the Codex-native infrastructure selectively:
plugins/clo-author/skills/plugins/clo-author/references/paper/, scripts/, data/, quality_reports/, and MEMORY.md