[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"project-75071":3},{"id":4,"name":5,"fullName":6,"owner":7,"repo":5,"description":8,"homepage":9,"htmlUrl":9,"language":10,"languages":9,"totalLinesOfCode":9,"stars":11,"forks":12,"watchers":13,"openIssues":14,"contributorsCount":15,"subscribersCount":15,"size":15,"stars1d":16,"stars7d":17,"stars30d":18,"stars90d":15,"forks30d":15,"starsTrendScore":19,"compositeScore":20,"rankGlobal":9,"rankLanguage":9,"license":21,"archived":22,"fork":22,"defaultBranch":23,"hasWiki":24,"hasPages":24,"topics":25,"createdAt":9,"pushedAt":9,"updatedAt":26,"readmeContent":27,"aiSummary":28,"trendingCount":15,"starSnapshotCount":15,"syncStatus":29,"lastSyncTime":30,"discoverSource":31},75071,"claude-code-my-workflow","pedrohcgs\u002Fclaude-code-my-workflow","pedrohcgs","A ready-to-fork Claude Code template for academics using LaTeX\u002FBeamer + R. Multi-agent review, quality gates, adversarial QA, and replication protocols.",null,"HTML",1246,2526,7,8,0,26,60,164,78,108,"MIT License",false,"main",true,[],"2026-06-12 04:01:17","# My Claude Code Setup\n\n[![License: MIT](https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FLicense-MIT-blue.svg)](LICENSE)\n[![Changelog](https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FSee-CHANGELOG-blue.svg)](CHANGELOG.md)\n[![Contributing](https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FPRs-welcome-brightgreen.svg)](.github\u002FCONTRIBUTING.md)\n\n> **Actively maintained.** A summary of how I use Claude Code for academic work — slides, papers, data analysis, and more — packaged so you can fork it for your own research. See [CHANGELOG.md](CHANGELOG.md) for the latest changes.\n\n**Live site:** [psantanna.com\u002Fclaude-code-my-workflow](https:\u002F\u002Fpsantanna.com\u002Fclaude-code-my-workflow\u002F)\n\nA ready-to-fork foundation for AI-assisted academic work. You describe what you want — lecture slides, a research paper, a data analysis, a replication package — and Claude plans the approach, runs specialized agents, fixes issues, verifies quality, and presents results. Like a contractor who handles the entire job. Extracted from a production PhD course and extended by a growing [community](#community--extensions).\n\n---\n\n## Quick Start (5–10 minutes, plus ~30 min for first-time installs)\n\n> **Before you start:** Claude Code + git are the minimum. To run the included `HelloWorld` demos end-to-end you also need XeLaTeX (Beamer sample) and Quarto (Quarto sample). R and the GitHub CLI are recommended. Python 3 is used by a few internal scripts (`check-palette-sync.py`, `check-tikz-prevention.py`) and is pre-installed on macOS\u002FLinux. Full list in [Prerequisites](#prerequisites) below. Fastest path: clone first, then run `.\u002Fscripts\u002Fvalidate-setup.sh` — it reports exactly what's missing with install links.\n>\n> **Only need Python\u002FR\u002Fmarkdown?** You don't need XeLaTeX or Quarto. The agents, rules, skills, and orchestration patterns work for any text\u002Fcode artifact. Skip the `HelloWorld` demos and head straight to `\u002Fdata-analysis`, `\u002Freview-paper`, `\u002Flit-review`, or `\u002Freview-r`.\n>\n> **Session 2 onwards:** [MEMORY.md](MEMORY.md) (committed) collects generic `[LEARN]` entries that help all forkers; `.claude\u002Fstate\u002Fpersonal-memory.md` (gitignored) is for machine-specific notes. See [`.claude\u002Frules\u002Fmeta-governance.md`](.claude\u002Frules\u002Fmeta-governance.md) for the distinction.\n\n### 1. Fork & Clone\n\n```bash\n# Fork this repo on GitHub (click \"Fork\" on the repo page), then:\ngit clone https:\u002F\u002Fgithub.com\u002FYOUR_USERNAME\u002Fclaude-code-my-workflow.git my-project\ncd my-project\n.\u002Fscripts\u002Fvalidate-setup.sh        # reports missing tools with install links\n```\n\nReplace `YOUR_USERNAME` with your GitHub username.\n\n### 2. Start Claude Code and Paste This Prompt\n\n```bash\nclaude\n```\n\n**Using VS Code?** Open the Claude Code panel instead. Everything works the same — see the [full guide](https:\u002F\u002Fpsantanna.com\u002Fclaude-code-my-workflow\u002Fworkflow-guide.html#sec-setup) for details.\n\n> **Avoid prompt fatigue.** Out of the box, Claude Code asks permission for every tool invocation. After the first few approvals, toggle **Auto-accept edits** mode (a keybinding; see the [permission modes section](https:\u002F\u002Fpsantanna.com\u002Fclaude-code-my-workflow\u002Fworkflow-guide.html#settings---permissions-and-hooks) of the guide) or run `claude --permission-mode acceptEdits`. For fully-autonomous runs on a trusted repo, **Bypass** mode skips prompts entirely. The template's `.claude\u002Fsettings.json` pre-approves ~100 common Bash and Edit\u002FWrite patterns, so even at default permissions most work is unattended.\n\nThen paste the [starter prompt](https:\u002F\u002Fpsantanna.com\u002Fclaude-code-my-workflow\u002Fworkflow-guide.html#sec-first-session) from the guide, filling in your project details:\n\n> I am starting to work on **[PROJECT NAME]** in this repo. **[Describe your project in 2–3 sentences.]** I've set up the Claude Code academic workflow... Please read the configuration files and adapt them for my project. Enter plan mode and start.\n\nThe [full guide](https:\u002F\u002Fpsantanna.com\u002Fclaude-code-my-workflow\u002Fworkflow-guide.html#sec-first-session) has the complete starter prompt with all the details.\n\n**What this does:** Claude reads all the configuration files, fills in your project name, institution, and preferences, then enters contractor mode — planning, implementing, and (within the skill you invoke) running the review + verify loop. You approve the plan, invoke a skill, and the skill handles the rest within its scope.\n\n> **Heavily adapting CLAUDE.md for a non-academic project?** Anthropic's built-in `\u002Finit` command will re-derive a `CLAUDE.md` from your codebase as a starting point. The pre-shipped CLAUDE.md in this template already covers the academic setup — you only need `\u002Finit` if your fork diverges substantially (e.g., a Python\u002FML project that doesn't use LaTeX or Quarto).\n\n### 3. Verify Your Setup\n\nBefore building real lectures, confirm your environment works:\n\n```bash\n.\u002Fscripts\u002Fvalidate-setup.sh        # Checks XeLaTeX, Quarto, Python, git, etc.\n```\n\nThen inside Claude:\n\n```text\n\u002Fcompile-latex HelloWorld          # Compiles Slides\u002FHelloWorld.tex to PDF\n\u002Fdeploy HelloWorld                 # Renders Quarto\u002FHelloWorld.qmd to HTML\n```\n\nIf both succeed, delete `Slides\u002FHelloWorld.tex` and `Quarto\u002FHelloWorld.qmd` and start on your real work.\n\n---\n\n## How It Works\n\n### Contractor Mode\n\nYou describe a task. For complex or ambiguous requests, Claude first creates a requirements specification with MUST\u002FSHOULD\u002FMAY priorities and clarity status (CLEAR\u002FASSUMED\u002FBLOCKED). You approve the spec, then Claude plans the approach and invokes the right skill (e.g. `\u002Fcreate-lecture`, `\u002Fqa-quarto`, `\u002Freview-paper --adversarial`). That skill implements the orchestrator pattern internally — implement, verify, review, fix, re-verify, score — and returns a summary when the work meets quality standards. Say \"just do it\" and it auto-commits when the score clears 80.\n\n### Specialized Agents\n\nInstead of one general-purpose reviewer, 14 focused agents each check one dimension. A representative sample:\n\n- **proofreader** — grammar\u002Ftypos\n- **slide-auditor** — visual layout\n- **pedagogy-reviewer** — teaching quality\n- **r-reviewer** — R code quality\n- **domain-reviewer** — field-specific correctness, slides (template — customize for your field)\n- **domain-referee** \u002F **methods-referee** \u002F **editor** — manuscript peer-review pipeline (`\u002Freview-paper --peer`)\n\nEach is better at its narrow task than a generalist would be. The `\u002Fslide-excellence` skill runs the slide-review agents in parallel; `\u002Freview-paper --peer` runs the paper-review pipeline. The same pattern extends to any academic artifact — manuscripts, data pipelines, proposals.\n\n### Adversarial QA\n\nTwo agents work in opposition: the **critic** reads both Beamer and Quarto and produces harsh findings. The **fixer** implements exactly what the critic found. They loop until the critic says \"APPROVED\" (or 5 rounds max). This catches errors that single-pass review misses.\n\n### Quality Review\n\nEvery artifact gets a score (0–100). Scores below threshold halt the workflow and surface the findings — the user decides whether to fix or explicitly override:\n\n- **80** — commit threshold\n- **90** — PR threshold\n- **95** — excellence (aspirational)\n\n> **Framing honesty:** Thresholds are advisory at the harness level — the `\u002Fcommit` skill runs quality checks and halts on failure, but there is no pre-commit git hook that blocks a direct `git commit`. If you bypass the skill, you bypass the review. For hard enforcement, configure a git pre-commit hook.\n\n### Context Survival\n\nPlans, specifications, and session logs survive auto-compression and session boundaries. The PreCompact hook saves a context snapshot before Claude's auto-compression triggers, ensuring critical decisions are never lost. MEMORY.md accumulates learning across sessions, so patterns discovered in one session inform future work.\n\n---\n\n## The Guide\n\nFor a comprehensive walkthrough, read the **[full guide](https:\u002F\u002Fpsantanna.com\u002Fclaude-code-my-workflow\u002Fworkflow-guide.html)** (or see the [source](guide\u002Fworkflow-guide.qmd)).\n\nIt covers:\n1. **Why This Workflow Exists** — the problem and the vision\n2. **Getting Started** — fork, paste one prompt, and Claude sets up the rest\n3. **The System in Action** — specialized agents, adversarial QA, quality scoring\n4. **The Building Blocks** — CLAUDE.md, rules, skills, agents, hooks, memory\n5. **Workflow Patterns** — slides, research, reproducibility, presentation rhetoric, sequential adversarial audits, and more\n6. **The Ecosystem** — extensions by clo-author, claudeblattman, MixtapeTools, autoresearch, ClaudeCodeTools, and a growing community\n7. **Customizing for Your Domain** — creating your own reviewers and knowledge bases\n\n### 2026 Features\n\nThe guide covers Claude Code's latest capabilities:\n\n- **Effort levels** — `\u002Feffort` command for cost vs. thoroughness tradeoffs (low\u002Fmedium\u002Fhigh\u002Fmax)\n- **Skill frontmatter** — `effort`, `context: fork`, `agent`, `hooks`, and dynamic content (`$ARGUMENTS`, `!command` syntax)\n- **Permission modes** — Normal, Auto-accept, Plan, Bypass for different workflows\n- **Hook handler types** — command, prompt, and HTTP handlers with 20+ hook events\n- **Advanced agent configuration** — model, maxTurns, isolation, tool restrictions\n- **Built-in skills** — `\u002Fbatch` for parallel refactoring, `\u002Fsimplify` for code review, `\u002Fremote-control` for browser bridge\n- **Plugins** — `\u002Fdiscover-plugins` for third-party extensions\n\n---\n\n## Use Cases\n\n| Academic Task | How This Workflow Helps |\n|---------------|----------------------|\n| Lecture slides (Beamer\u002FQuarto) | Full creation, translation, multi-agent review, deployment |\n| Research papers | Literature review, manuscript review, simulated peer review |\n| Data analysis | End-to-end R pipelines, replication verification, publication-ready output |\n| Replication packages | AEA-compliant packaging, reproducibility audit trails |\n| Presentations | Rhetoric of decks principles, visual audit, cognitive load review |\n| Research proposals | Structured drafting with adversarial critique |\n\n---\n\n## What's Included\n\n\u003Cdetails>\n\u003Csummary>\u003Cstrong>14 agents, 30 skills, 24 rules, 6 hooks\u003C\u002Fstrong> (click to expand)\u003C\u002Fsummary>\n\n### Agents (`.claude\u002Fagents\u002F`)\n\n| Agent | What It Does |\n|-------|-------------|\n| `proofreader` | Grammar, typos, overflow, consistency review |\n| `slide-auditor` | Visual layout audit (overflow, font consistency, spacing) |\n| `pedagogy-reviewer` | 13-pattern pedagogical review (narrative arc, notation density, pacing) |\n| `r-reviewer` | R code quality, reproducibility, and domain correctness |\n| `tikz-reviewer` | Merciless TikZ diagram visual critique |\n| `beamer-translator` | Beamer-to-Quarto translation specialist |\n| `quarto-critic` | Adversarial QA comparing Quarto against Beamer benchmark |\n| `quarto-fixer` | Implements fixes from the critic agent |\n| `verifier` | End-to-end task completion verification |\n| `domain-reviewer` | **Template** for your field-specific substance reviewer |\n\n### Skills (`.claude\u002Fskills\u002F`)\n\n| Skill | What It Does |\n|-------|-------------|\n| `\u002Fcompile-latex` | 3-pass XeLaTeX compilation with bibtex |\n| `\u002Fdeploy` | Render Quarto + sync to GitHub Pages |\n| `\u002Fextract-tikz` | TikZ diagrams to PDF to SVG pipeline |\n| `\u002Fproofread` | Launch proofreader on a file |\n| `\u002Fvisual-audit` | Launch slide-auditor on a file |\n| `\u002Fpedagogy-review` | Launch pedagogy-reviewer on a file |\n| `\u002Freview-r` | Launch R code reviewer |\n| `\u002Fqa-quarto` | Adversarial critic-fixer loop (max 5 rounds) |\n| `\u002Fslide-excellence` | Combined multi-agent review |\n| `\u002Ftranslate-to-quarto` | Full 11-phase Beamer-to-Quarto translation |\n| `\u002Fvalidate-bib` | Cross-reference citations against bibliography |\n| `\u002Fdevils-advocate` | Challenge design decisions before committing |\n| `\u002Fcreate-lecture` | Full lecture creation workflow |\n| `\u002Fcommit` | Stage, commit, create PR, and merge to main |\n| `\u002Flit-review` | Literature search, synthesis, and gap identification |\n| `\u002Fresearch-ideation` | Generate research questions and empirical strategies |\n| `\u002Finterview-me` | Interactive interview to formalize a research idea |\n| `\u002Freview-paper` | Manuscript review: structure, econometrics, referee objections |\n| `\u002Fdata-analysis` | End-to-end R analysis with publication-ready output |\n| `\u002Flearn` | Extract non-obvious discoveries into persistent skills |\n| `\u002Fcontext-status` | Show session health and context usage |\n| `\u002Fdeep-audit` | Repository-wide consistency audit |\n| `\u002Fpermission-check` | Diagnose permission layers when prompts fire unexpectedly |\n| `\u002Faudit-reproducibility` | Enforce tolerance thresholds on paper ↔ code numeric claims |\n| `\u002Fnew-diagram` | Scaffold a TikZ diagram from the snippet gallery with prevention + review |\n| `\u002Frespond-to-referees` | R&R response-letter generator (maps referee comments to revisions) |\n| `\u002Fseven-pass-review` | Seven-pass adversarial manuscript review (parallel forked subagents) |\n| `\u002Fcheckpoint` | Structured session-handoff snapshot (state + plan pointers + next actions). Companion to narrative session logs. |\n| `\u002Fpreregister` | Generate a preregistration document (OSF \u002F AsPredicted \u002F AEA RCT Registry style) from a research spec |\n\n### Research Workflow\n\n| Feature | What It Does |\n|---------|-------------|\n| Exploration folder | Structured `explorations\u002F` sandbox with graduate\u002Farchive lifecycle |\n| Fast-track workflow | 60\u002F100 quality threshold for rapid prototyping |\n| Simplified orchestrator | implement → verify → score → done (no multi-round reviews) |\n| Enhanced session logging | Structured tables for changes, decisions, verification |\n| Merge-only reporting | Quality reports at merge time only |\n| Math line-length exception | Long lines acceptable for documented formulas |\n| Workflow quick reference | One-page cheat sheet at `.claude\u002FWORKFLOW_QUICK_REF.md` |\n\n### Rules (`.claude\u002Frules\u002F`)\n\nRules use path-scoped loading: **always-on** rules load every session (~100 lines total); **path-scoped** rules load only when Claude works on matching files. Claude follows ~150 instructions reliably, so less is more.\n\n**Always-on** (no `paths:` frontmatter — load every session):\n\n| Rule | What It Enforces |\n|------|-----------------|\n| `plan-first-workflow` | Plan mode for non-trivial tasks + context preservation |\n| `orchestrator-protocol` | Contractor mode: implement → verify → review → fix → score |\n| `session-logging` | Three logging triggers: post-plan, incremental, end-of-session |\n| `meta-governance` | Template vs. working project distinctions |\n\n**Path-scoped** (load only when working on matching files):\n\n| Rule | Triggers On | What It Enforces |\n|------|------------|-----------------|\n| `verification-protocol` | `.tex`, `.qmd`, `docs\u002F` | Task completion checklist |\n| `single-source-of-truth` | `Figures\u002F`, `.tex`, `.qmd` | No content duplication; Beamer is authoritative |\n| `quality-gates` | `.tex`, `.qmd`, `*.R` | 80\u002F90\u002F95 scoring + tolerance thresholds |\n| `r-code-conventions` | `*.R` | R coding standards + math line-length exception |\n| `tikz-visual-quality` | `.tex` | TikZ diagram visual standards |\n| `beamer-quarto-sync` | `.tex`, `.qmd` | Auto-sync Beamer edits to Quarto |\n| `pdf-processing` | `master_supporting_docs\u002F` | Safe large PDF handling |\n| `proofreading-protocol` | `.tex`, `.qmd`, `quality_reports\u002F` | Propose-first, then apply with approval |\n| `no-pause-beamer` | `.tex` | No overlay commands in Beamer |\n| `replication-protocol` | `*.R` | Replicate original results before extending |\n| `knowledge-base-template` | `.tex`, `.qmd`, `*.R` | Notation\u002Fapplication registry template |\n| `orchestrator-research` | `*.R`, `explorations\u002F` | Simple orchestrator for research (no multi-round reviews) |\n| `exploration-folder-protocol` | `explorations\u002F` | Structured sandbox for experimental work |\n| `exploration-fast-track` | `explorations\u002F` | Lightweight exploration workflow (60\u002F100 threshold) |\n\n### Templates (`templates\u002F`)\n\n| Template | What It Does |\n|----------|-------------|\n| `session-log.md` | Structured session logging format |\n| `quality-report.md` | Merge-time quality report format |\n| `exploration-readme.md` | Exploration project README template |\n| `archive-readme.md` | Archive documentation template |\n| `requirements-spec.md` | MUST\u002FSHOULD\u002FMAY requirements framework with clarity status |\n| `constitutional-governance.md` | Template for defining non-negotiable principles vs. preferences |\n| `skill-template.md` | Academic skill creation template with domain-specific examples |\n\n\u003C\u002Fdetails>\n\n---\n\n## Prerequisites\n\n| Tool | Required For | Install |\n|------|-------------|---------|\n| [Claude Code](https:\u002F\u002Fcode.claude.com\u002Fdocs\u002Fen\u002Foverview) | Everything | [claude.ai\u002Finstall](https:\u002F\u002Fclaude.ai\u002Finstall) |\n| git | Clone + version control | [git-scm.com](https:\u002F\u002Fgit-scm.com\u002Fdownloads) |\n| Python 3 (3.9+) | Internal checkers (palette sync, TikZ prevention) | Preinstalled on macOS\u002FLinux; [python.org](https:\u002F\u002Fwww.python.org\u002F) for Windows |\n| XeLaTeX | LaTeX compilation (Beamer `HelloWorld`, real lectures) | [TeX Live](https:\u002F\u002Ftug.org\u002Ftexlive\u002F) or [MacTeX](https:\u002F\u002Ftug.org\u002Fmactex\u002F) |\n| [Quarto](https:\u002F\u002Fquarto.org) | Web slides (Quarto `HelloWorld`, real lectures) | [quarto.org\u002Fdocs\u002Fget-started](https:\u002F\u002Fquarto.org\u002Fdocs\u002Fget-started\u002F) |\n| R | Figures and analysis (`\u002Fdata-analysis`, `scripts\u002FR\u002F` template) | [r-project.org](https:\u002F\u002Fwww.r-project.org\u002F) |\n| pdf2svg | TikZ → SVG for Quarto (`\u002Fextract-tikz`) | `brew install pdf2svg` (macOS), `apt install pdf2svg` (Debian) |\n| [gh CLI](https:\u002F\u002Fcli.github.com\u002F) | PR \u002F issue workflow | `brew install gh` (macOS), `apt install gh` (Debian) |\n\n**Minimum to fork this template:** Claude Code + git + Python 3 (Python is already installed on macOS\u002FLinux).\n\n**Minimum to run the included HelloWorld demos end-to-end:** add XeLaTeX (for `\u002Fcompile-latex HelloWorld`) and Quarto (for `\u002Fdeploy HelloWorld`).\n\n**Your real lectures may need more** — R for `scripts\u002FR\u002F` analyses, pdf2svg if you use TikZ extraction, gh CLI if you use the PR-based commit workflow. `.\u002Fscripts\u002Fvalidate-setup.sh` reports which of these are installed and what each unlocks.\n\n---\n\n## Adapting for Your Field\n\n1. **Fill in the knowledge base** (`.claude\u002Frules\u002Fknowledge-base-template.md`) with your notation, applications, and design principles\n2. **Customize the domain reviewer** (`.claude\u002Fagents\u002Fdomain-reviewer.md`) with review lenses specific to your field\n3. **Update the color palette** — this is a **two-surface contract**: change the HEX values at the top of **both** [`Preambles\u002Fheader.tex`](Preambles\u002Fheader.tex) (Beamer\u002FTikZ) **and** [`Quarto\u002Ftheme-template.scss`](Quarto\u002Ftheme-template.scss) (Quarto slides) so they agree. Then run `.\u002Fscripts\u002Fcheck-palette-sync.sh` to verify. Forgetting one surface silently produces mismatched Beamer vs. Quarto renderings. See [`Preambles\u002FREADME.md`](Preambles\u002FREADME.md) for the full contract and the TikZ style library.\n4. **Add field-specific R pitfalls** to `.claude\u002Frules\u002Fr-code-conventions.md`\n5. **Fill in the lecture mapping** in `.claude\u002Frules\u002Fbeamer-quarto-sync.md`\n6. **Customize the workflow quick reference** (`.claude\u002FWORKFLOW_QUICK_REF.md`) with your non-negotiables and preferences\n7. **Set up the exploration folder** (`explorations\u002F`) for experimental work\n\n---\n\n## Additional Resources\n\n- [Claude Code Documentation](https:\u002F\u002Fcode.claude.com\u002Fdocs\u002Fen\u002Foverview)\n- [Writing a Good CLAUDE.md](https:\u002F\u002Fcode.claude.com\u002Fdocs\u002Fen\u002Fmemory) — official guidance on project memory\n\n---\n\n## Origin\n\nThis infrastructure was extracted from **Econ 730: Causal Panel Data** at Emory University, developed by Pedro Sant'Anna using Claude Code over 6+ sessions. The course produced 6 complete PhD lecture decks with 800+ slides, interactive Quarto versions with plotly charts, and full R replication packages — all managed through this multi-agent workflow. The patterns are domain-agnostic: the same agents, rules, and orchestrator work for any academic project.\n\n---\n\n## Community & Extensions\n\nAs of March 2026, **15+ research groups** across economics, energy, political science, and engineering have forked and adapted this workflow. The infrastructure (orchestrator, hooks, quality gates) transfers without modification.\n\n**Extended workflows:**\n\n- **[clo-author](https:\u002F\u002Fgithub.com\u002Fhugosantanna\u002Fclo-author)** by Hugo Sant'Anna (UAB) — Paper-centric research workflows with 17 specialized agents (6 worker-critic pairs plus referees, data-engineer, verifier), simulated blind peer review, AEA replication compliance, and full research lifecycle management. **The `\u002Freview-paper --peer \u003Cjournal>` pipeline in this template is adapted from clo-author with Hugo's permission** (pipeline shape, 6-way disposition taxonomy, journal-calibration schema, paper-type branching). Thanks, Hugo.\n- **[claudeblattman](https:\u002F\u002Fgithub.com\u002Fchrisblattman\u002Fclaudeblattman)** by Chris Blattman (U Chicago) — Comprehensive guide for non-technical academics: executive assistant workflows, proposal writing, agent debates, and self-improving configuration\n- **[MixtapeTools](https:\u002F\u002Fgithub.com\u002Fscunning1975\u002FMixtapeTools)** by Scott Cunningham (Baylor) — The Rhetoric of Decks: philosophy and practice of beautiful, rhetorically effective academic presentations\n- **[autoresearch](https:\u002F\u002Fgithub.com\u002Fkarpathy\u002Fautoresearch)** by Andrej Karpathy — Constraint-based autonomous research with `program.md` as constitutional document\n- **[ClaudeCodeTools](https:\u002F\u002Fgithub.com\u002Faspi6246\u002FClaudeCodeTools)** — \"The Editor\" persona: seven-audit sequential paper review protocol\n\nSee the [guide's ecosystem section](https:\u002F\u002Fpsantanna.com\u002Fclaude-code-my-workflow\u002Fworkflow-guide.html#sec-ecosystem) for detailed descriptions, design principles, and more resources.\n\n---\n\n## Versioning & Contributing\n\n- **What's new:** see [CHANGELOG.md](CHANGELOG.md). We follow loose semver — breaking changes get major bumps so you can decide when to pull updates.\n- **How to contribute:** see [.github\u002FCONTRIBUTING.md](.github\u002FCONTRIBUTING.md). PRs welcome for generalizable improvements; fork-specific work stays in your fork.\n- **Pin to a version:** `git checkout v1.8.0` (current as of 2026-04-27).\n\n---\n\n## License\n\nMIT License. See [LICENSE](LICENSE).\n","该项目提供了一个基于Claude Code的模板，适用于使用LaTeX\u002FBeamer和R进行学术工作的研究人员。其核心功能包括多代理审核、质量控制门、对抗性QA以及复制协议，旨在自动化从计划到结果呈现的整个研究流程。技术特点在于通过AI辅助的方式，帮助用户生成论文、报告、数据分析等学术材料，并确保内容的质量与一致性。适合需要提高科研效率，尤其是希望在撰写文档、准备演示文稿或执行复杂数据处理任务时获得支持的研究人员使用。",2,"2026-06-11 03:52:12","high_star"]