Michael McBride
Community@michaelsengineering · McLean, VA
AI | ML Research Engineer & Founder
Agent Skills by Michael McBride
Showing 61 vetted skills indexed across 2 GitHub repositories.
MichaelsEngineeringgpd-respond-to-referees
Structure point-by-point responses to referee reports and revise LaTeX manuscripts accordingly.
MichaelsEngineeringgpd-error-propagation
Propagates uncertainties through multi-step physics derivation chains to produce ranked error budgets.
MichaelsEngineeringgpd-insert-phase
Insert urgent research work as decimal phases between existing roadmap phases.
MichaelsEngineeringgpd-branch-hypothesis
Creates a git hypothesis branch with forked research state and hypothesis documentation.
MichaelsEngineeringgpd-new-milestone
Initializes a new research milestone cycle by updating PROJECT.md, defining requirements, and generating a roadmap.
MichaelsEngineeringgpd-error-patterns
Displays accumulated physics error patterns from the project error database with optional category filtering.
MichaelsEngineeringgpd-map-research
Maps physics research projects into seven structured documents using parallel mapper agents.
MichaelsEngineeringgpd-explain
Generates rigorous physics concept explanations with verified citations and project context.
MichaelsEngineeringgpd-audit-milestone
Audit research milestone completion against phase verifications and original research goals.
MichaelsEngineeringgpd-help
Displays the complete GPD command reference and state-aware usage guidance.
MichaelsEngineeringgpd-compare-results
Compares internal results, baselines, or methods and emits machine-readable verdict artifacts.
MichaelsEngineeringgpd-regression-check
Re-verify previously verified physics claims across project phases to detect regressions.
MichaelsEngineeringgpd-reapply-patches
Reapply backed-up local modifications to GPD files after an update reinstall.
MichaelsEngineeringgpd-verify-work
Verify research results through physics consistency checks and computational spot-checks.
MichaelsEngineeringgpd-settings
Configure GPD workflow toggles, tier model overrides, review cadence, and git branching preferences.
MichaelsEngineeringgpd-sync-state
Reconcile diverged STATE.md and state.json files after manual edits or corruption.
MichaelsEngineeringgpd-check-todos
Lists pending research todos and routes a selected todo to an action.
MichaelsEngineeringgpd-sensitivity-analysis
Computes parameter sensitivity coefficients and ranks uncertainty contributions in physics calculations.
MichaelsEngineeringgpd-parameter-sweep
Executes systematic parameter sweeps with wave-based parallel execution and result aggregation.
MichaelsEngineeringgpd-export
Export research results to HTML, LaTeX, or ZIP reproducibility packages.
MichaelsEngineeringgpd-record-insight
Record project-specific learnings and error patterns to a structured insights ledger.
MichaelsEngineeringgpd-list-phase-assumptions
Surfaces the AI's assumptions about a research phase before planning begins.
MichaelsEngineeringgpd-discover
Runs discovery investigations of physics methods and literature to produce RESEARCH.md before phase planning.
MichaelsEngineeringgpd-debug
Diagnose physics calculation errors through parallel subagent investigation with persistent debug sessions.
Frequently Asked Questions About Michael McBride
FAQPage SchemaWhat tasks can I accomplish with MichaelsEngineering's GPD skills?▼
The 62 GPD skills cover the full physics research lifecycle: project initialization, roadmap and milestone planning, phase execution, equation derivation, sensitivity analysis, parameter sweeps, error propagation, regression checks, hypothesis branching, literature reviews, paper writing, six-stage peer review, referee responses, and arXiv submission packaging.
Who are the GPD skills designed for?▼
They target physics researchers, computational scientists, and ML research engineers who need structured, auditable research projects. Skills enforce verification contracts, convention consistency, and decision logging, suiting graduate students through principal investigators managing multi-phase theoretical or numerical investigations.
How do GPD skills manage project state and execution?▼
GPD stores state in .gpd/ artifacts including STATE.md, state.json, ROADMAP.md, and DECISIONS.md. Skills like gpd-sync-state reconcile diverged files, gpd-resume-work restores context, gpd-execute-phase runs plans in waves with checkpoints, and gpd-undo reverts the last commit with a safety checkpoint.
What prerequisites do GPD skills require?▼
Most skills are project-required, needing .gpd/ROADMAP.md and STATE.md initialized via gpd-new-project. Publication skills (peer-review, write-paper, arxiv-submission) require an existing LaTeX manuscript under paper/, manuscript/, or draft/ plus verification reports and bibliography audits before running.
How does GPD ensure scientific correctness of results?▼
Verification skills audit dimensional consistency, enumerate and test limiting cases, run numerical convergence refinement studies, propagate uncertainties into ERROR-BUDGET.md, compute dimensionless sensitivities, and re-validate previously verified claims via regression checks, emitting VERIFICATION.md and COMPARISON.md evidence artifacts.