Lowell Mason
Community@lowmason
Senior data scientist. Aspiring Bayesian.
Agent Skills by Lowell Mason
Showing 21 vetted skills indexed across 1 GitHub repositories.
lowmasonusing-git-worktrees
Set up isolated workspaces using native tooling or git worktree fallback.
lowmasonvalidate-data
Validate datasets and analyses for publication readiness with Polars.
lowmasontest-driven-development
Enforce a test-first workflow with failing tests before production code.
lowmasonsystematic-debugging
Guide root-cause investigation and phase-driven analysis for software bugs.
lowmasonbls-data-context
Summarize BLS labor statistics programs and cross-program data concepts.
lowmasondispatching-parallel-agents
Dispatch independent agents to investigate separate failures in parallel.
lowmasonexplore-data
Profile Parquet or CSV datasets with Polars to surface data quality issues.
lowmasontech-debt
Triage technical debt signals in Polars, NumPyro, PyMC, and BLS codebases into a prioritized backlog.
lowmasonexecuting-plans
Execute written plans stepwise with status tracking and verifications.
lowmasonfinishing-a-development-branch
Guide merge, PR, or cleanup decisions after verifying tests in Git workflows.
lowmasonbrainstorming
Convert raw ideas into polished design specifications with structured prompts.
lowmasonwriting-plans
Transform vague specs into structured task plans with testable criteria.
lowmasonbayesian-workflow
Scaffold Bayesian workflows with NumPyro and ArviZ diagnostics.
lowmasonrequesting-code-review
Dispatches a subagent to review code against plan, quality, and architecture criteria.
lowmasonreceiving-code-review
Evaluate code review feedback with verification against codebase requirements.
lowmasonwriting-skills
Standardize SKILL.md creation with RED-GREEN-REFACTOR testing workflows.
lowmasondevelop-testing-strategy
Plan invariant-based testing strategies for data-science codebases.
lowmasonverification-before-completion
Enforce evidence-based verification before declaring task completion.
lowmasonsubagent-driven-development
Dispatch fresh subagents per task with per-task reviews and a final aggregation pass.
lowmasondesign-architecture
Create and evaluate Architecture Decision Records for cross-cutting data design decisions.
lowmasonrecommend-probabilistic-model
Diagnose modeling problems and map data signals to probabilistic-method families with Murphy §refs and pyprobml notebooks.