Claude Code My Workflow
A ready-to-fork Claude Code template for academic work: slides, papers and data analysis with multi-agent review and quality gates
High risk
We rate an entry high when the tool writes to external systems, handles money, production databases or secrets, or runs arbitrary commands. The CLI installs it only with your consent.
Why this level
- Ships bypassPermissions mode by default with broad Bash(*), Edit(**) and Write(**) allows, prompting for almost nothing
- Runs scripts, hooks and pre-commit gates that execute code and drive compilation and git
Install
Manual install
git clone https://github.com/pedrohcgs/claude-code-my-workflow.git my-project
cd my-project
./scripts/validate-setup.shFirst fork the repository on GitHub and clone your fork. The validate-setup.sh script checks XeLaTeX, Quarto, Python and git and points to install links.
This is third-party code. Review the repository files before installing.
What it does
This is a repository template that you fork for your own research work. It ships about 60 skills, 18 agents, and a set of rules, hooks and scripts. You state a goal, the agents do the work, and quality gates decide when it is good enough: one backtest.sh command runs about ten checks, and a pre-commit hook requires a score of at least 80. There is an adversarial mode where a critic finds issues and a fixer resolves them in a loop, plus narrow reviewers such as a proofreader, an R code reviewer and a slide pedagogy reviewer. Skills cover common tasks: building a lecture, reviewing a paper, analyzing data, auditing reproducibility. The template is tuned for LaTeX, Beamer, Quarto and R in economics and political science, but the patterns work for any text or code.
Who it is for. For researchers and teachers who run papers, slides and data analysis in Claude Code.
Good fit when
- You need a research repo with review and quality gates out of the box
- You need to write or review a paper, a lecture or a data analysis with multi-agent checks
- You need a reproducibility audit so the numbers in the text match the script output
Not a fit when
- You do not work in Claude Code
- You need a lightweight setup without the LaTeX and R toolchain and strict quality gates
Example request
I am starting work on a paper in this repo, read the configuration, enter plan mode and propose a planLimitations
You need Claude Code and git, plus XeLaTeX and Quarto to run the demos end to end, ideally R and the GitHub CLI, and Python for the checks. The template is heavily tuned for economics and political science and needs adapting for another field. By default .claude/settings.json sets bypassPermissions mode with broad Bash(*), Edit(**) and Write(**) allows, so it prompts for almost nothing; this is a deliberate power-user choice and is worth tightening before use.
How to disable. It is your own fork, so simply delete the cloned repository. To remove only the automation, drop the .claude folder with rules and hooks and disable the pre-commit hook from .githooks.
Security check
- Ships bypassPermissions mode by default with broad Bash(*), Edit(**) and Write(**) allows, prompting for almost nothing
- Runs scripts, hooks and pre-commit gates that execute code and drive compilation and git
README in short
The README presents a forkable Claude Code template for academic work, extracted from a PhD course and extended by the community. It bundles about 60 skills, 18 agents, 37 rules and 8 hooks, a contractor mode, an adversarial critic and fixer loop, and a quality score up to 100 with thresholds of 80 to commit and 90 for a PR. The backtest.sh command runs about ten gates, and a pre-commit hook keeps the check from being bypassed. Install is by forking and cloning, with validate-setup.sh reporting what is missing. It is tuned for LaTeX, Beamer, Quarto and R, MIT licensed.
FAQ
Is it a plugin or a standalone repository?
It is a repository template that you fork and clone. Skills, agents and rules live in the .claude folder, and nothing installs through /plugin.
Does it work outside economics?
The core patterns, agents and quality gates work for any text or code. But the journal profiles, reviewers and the LaTeX and R toolchain are tuned for economics and political science and need reworking for your field.
Related
Microsoft's utility and MCP server that turn PDF, Word, Excel, PowerPoint, HTML and other files into Markdown for models
Anthropic skills
Skills
Anthropic's official skill examples: docx, pdf, pptx and xlsx documents, design, MCP builder and the Agent Skills spec
A skill that builds editable PPTX decks from PDF, DOCX, URLs and Markdown with native shapes, charts and animations
A skill that draws editorial diagrams as standalone HTML and SVG files: architecture, flowcharts, ER, funnels and about forty more types