Fred Costa Milhome
Community@fredmilhome · São Paulo
Agent Skills by Fred Costa Milhome
Showing 52 vetted skills indexed across 1 GitHub repositories.
commit
Automate the commit-PR-merge workflow with pre-commit quality checks.
deep-audit
Audit repository infrastructure with parallel agents to resolve errors and inconsistencies.
learn
Extract session learnings into reusable YAML and Markdown skills.
checkpoint
Save and resume structured research state snapshots with active plans and decisions.
devils-advocate
Generate adversarial pedagogical questions from lecture documents to review slide decks.
preregister
Draft structured preregistration documents for research studies in OSF, AsPredicted, or AEA RCT Registry styles.
coauthor-brief
Generate a handoff brief summarizing git history, project artifacts, and environment configurations.
new-diagram
Generate TikZ diagrams with predefined snippets and validation rules.
did-event-study
Conduct difference-in-differences event study analysis with doubly robust estimators.
stata-replication
Build Stata replication pipelines with numbered .do files and stata-mcp server.
respond-to-eval
Analyze course evaluations into thematic clusters and generate teaching improvement plans.
submission-disclosures
Generate submission disclosure statements for manuscripts using journal policies.
deploy
Automate deployment of Quarto slides to GitHub Pages.
promote-memory
Review [LEARN] entries in personal-memory.md via five-critic majority vote.
teach-from-paper
Convert research papers into lecture outlines, slides, and discussion questions.
audit-reproducibility
Verify numeric claims in academic manuscripts against analysis outputs with tolerance thresholds.
translate-to-quarto
Convert LaTeX Beamer presentations to interactive Quarto RevealJS slides.
review-paper
Automate academic manuscript review with single-pass, adversarial, and peer-review modes.
replication-package
Create DCAS-compliant data and code replication packages for academic submissions.
qa-quarto
Compare Quarto HTML slides with Beamer PDFs for content and visual parity.
simulation-study
Run R Monte Carlo simulations to evaluate estimator bias and coverage.
context-status
Monitor session status and context usage within Claude Code.
diagnose
Diagnose root causes of empirical research code errors in R, Stata, and Python.
grant-proposal
Generates a draft research grant proposal from existing inputs.