Pedro H. C. Sant'Anna
Community@pedrohcgs
Agent Skills by Pedro H. C. Sant'Anna
Showing 59 vetted skills indexed across 1 GitHub repositories.
pedrohcgsdifferential-audit
Compare two implementations of the same computation across languages, versions, or replications.
pedrohcgsdeep-audit
Audits papers, proofs, and codebases via parallel adversarial finder and judge agents.
pedrohcgsadjudicate-review
Verify review findings against source material before applying fixes.
pedrohcgslearn
Extract reusable knowledge from sessions into persistent SKILL.md files.
pedrohcgscheckpoint
Writes structured session state snapshots to checkpoint files for cross-session handoff.
pedrohcgspreregister
Generates structured preregistration documents in OSF, AsPredicted, or AEA RCT Registry formats.
pedrohcgscoauthor-brief
Generate a Markdown handoff brief summarizing git deltas, artifact states, and reproduction steps for collaborators.
pedrohcgsnew-diagram
Scaffolds TikZ diagrams from snippet templates with prevention checks, compilation, and reviewer loops.
pedrohcgsstata-replication
Scaffolds and executes numbered Stata .do file pipelines for reproducible academic replication packages.
pedrohcgsblast-radius
Find and run every consumer before changing shared code interfaces.
pedrohcgsrespond-to-eval
Convert course evaluations into a classified teaching-improvement plan with syllabus and deck targets.
pedrohcgssubmission-disclosures
Generate journal-matched AI-use, CRediT, conflict-of-interest, and data-availability statements for manuscript submissions.
pedrohcgsdeploy
Render Quarto slides to HTML and sync them to docs for GitHub Pages deployment.
pedrohcgsoracle-review
Runs an external frontier-model referee on papers, proofs, and estimators, then triages findings.
pedrohcgspromote-memory
Reviews auto-memory learnings with a five-critic council to promote entries into MEMORY.md.
pedrohcgsteach-from-paper
Converts a research paper into a lecture outline, slide skeleton, discussion questions, and exercise brief.
pedrohcgsaudit-reproducibility
Cross-check numeric claims in manuscripts against R, Stata, and Python pipeline outputs.
pedrohcgsvoice-profile
Extract a personal writing voice profile from prior papers and audit drafts against it.
pedrohcgsreplication-package
Assemble a DCAS-compliant replication package with README, dataset manifest, and exhibit-to-script map.
pedrohcgsqa-quarto
Compares Quarto HTML slides against Beamer PDF benchmarks via iterative critic-fixer loops.
pedrohcgssimulation-study
Scaffold and run reproducible Monte Carlo simulation studies in R with MCSE reporting.
pedrohcgscontext-status
Reports context usage, active plan status, and session preservation state.
pedrohcgsdiagnose
Diagnose root causes of failing or wrong empirical results in research code.
pedrohcgsgrant-proposal
Scaffolds funder-shaped grant proposal drafts by composing existing research specs and delegated sections.