Muluh
Community@0xMuluh
Muluh provides disciplined R package engineering skills covering refactoring, API evolution, testing, dependency management, and Bioconductor submission readiness.
Agent Skills by Muluh
Showing 17 vetted skills indexed across 1 GitHub repositories.
0xMuluhmigrate-functionality-without-redesign
Migrate code functionality between packages while preserving existing behavior and parity.
0xMuluhdocument-a-real-workflow
Structures package documentation as question-driven workflows with minimal setup and interpreted results.
0xMuluhvalidate-a-package-change
Validates package changes through staged tests, builds, and failure attribution before review.
0xMuluhevolve-public-apis
Guides deliberate, compatibility-preserving changes to a package's public API contracts.
0xMuluhrefactor-without-behavioral-change
Guides code refactoring that preserves observable behavior with parity evidence and narrow structural scope.
0xMuluhprepare-a-bioconductor-ready-package
Audit and mature R packages for Bioconductor submission readiness and ecosystem interoperability.
0xMuluhfix-one-bug
Guides narrow bug fixes with regression tests and preserved surrounding behavior.
0xMuluhone-semantic-change
Enforces one semantic purpose per pull request through scope contracts and diff justification.
0xMuluhcore-development-philosophy
Guides software changes toward minimal conceptual scope and reuse of existing implementations.
0xMuluhremove-proven-dead-code
Removes unused code, branches, exports, and dependencies only after evidence proves non-use.
0xMuluhdeprecate-and-replace
Plan and execute API deprecations with canonical replacements, delegation, and migration guidance.
0xMuluhdesign-with-existing-representations
Guides selection of data representations in R packages using existing structures before creating new classes.
0xMuluhadd-a-thin-adapter
Design thin adapters that delegate to established upstream implementations.
0xMuluhreuse-before-implementing
Evaluates existing language, package, and ecosystem capabilities before implementing new code or dependencies.
0xMuluhtest-at-the-behavioral-boundary
Guides writing tests that assert public behavior contracts instead of private implementation details.
0xMuluhcharacterize-before-changing
Adds focused characterization tests that document existing code behavior before refactoring or bug fixes.
0xMuluhintroduce-dependencies-conservatively
Guides deliberate evaluation and management of package dependencies in software projects.
Frequently Asked Questions About Muluh
FAQPage SchemaWhat tasks can I accomplish using Muluh's skills?▼
You can refactor code without behavioral change, migrate functionality between modules, fix single bugs with regression tests, evolve public APIs safely, deprecate and replace old interfaces, remove proven dead code, manage dependencies conservatively, and prepare R packages for Bioconductor submission.
Who are Muluh's skills designed for?▼
They target R package developers and maintainers, especially those building or auditing Bioconductor packages. They also serve software engineers who want disciplined pull request scoping, characterization testing, and behavior-preserving refactoring practices.
How do the skills handle testing and validation?▼
Skills direct you to test at the behavioral boundary (inputs, outputs, classes, error conditions), add characterization tests before changing poorly understood code, and run a staged validation gate: focused evidence first, then broader tests, documentation builds, and package checks.
What is required before using the Bioconductor readiness skill?▼
You need an existing or planned R package. The skill then verifies current official Bioconductor devel requirements, reuses Bioconductor classes and public APIs, audits metadata, dependencies, vignettes, tests, data provenance, and requires passing R CMD build/check plus BiocCheck.
How do the skills approach dependencies and new code?▼
They require reuse before implementing: check the language, standard library, existing package public API, and current dependencies first. Dependency changes are treated as architectural decisions kept in their own semantic scope, never mixed with unrelated changes.