pi-skills

Guides machine-readable function catalogs for DuckDB extensions, including a `functions.yaml` source-of-truth pattern that can drive docs, wrappers, aliases, examples, and consistency checks. Use when an extension exposes many SQL functions or multiple wrapper surfaces.

Guides development in the DuckHTS repo, including DuckDB extension work, Rduckhts package integration, vendored htslib workflows, SQL and tinytest coverage, wasm/webR constraints, and exact-compatible rewrites of upstream genomics tools. Use when working in github.com/RGenomicsETL/duckhts.

Guides creation, maintenance, checking, testing, documenting, and releasing R packages using project-native workflows such as Makefiles, tinytest, roxygen2, base R, and optional tools like usethis, pkgdown, air, or jarl. Use when working on any R package or CRAN-style package workflow.

Guides cache, index, and annotation-store design for bioinformatics rewrites, including when to use mainstream interoperable formats versus specialized high-performance encodings. Use when startup cost, repeated lookup performance, interval access, or annotation storage strategy are central design concerns.

Guides library-first bioinformatics rewrites that expose mature native code through bindings, extensions, FFI, or embedded runtimes across R, Python, SQL, and wasm. Use when building reusable interfaces around existing C/C++ libraries instead of standalone-only rewrites.

Guides responsible AI-assisted rewriting and porting of bioinformatics tools, combining rewrites.bio principles with practical validation, attribution, compatibility, and maintenance discipline. Use when planning or implementing a rewrite, reimplementation, or high-performance port of an existing bioinformatics tool.

Guides design of bioinformatics kernels and rewrites that compute multiple validated outputs or statistics in one pass over the data, reducing repeated I/O and decompression while preserving explicit semantics. Use when designing fused readers, counters, coverage engines, or summary pipelines.

Guides DuckDB C extension API selection, deprecation handling, compatibility shims, and release policy. Use when balancing stable versus unstable APIs, managing breaking changes, or documenting deprecation strategy for SQL and wrapper surfaces.

Guides design of DuckDB extensions written primarily in C, including runtime ownership, function boundaries, state models, background services, concurrency, and separation of stable logic from volatile adapter code. Use when planning or restructuring a native DuckDB extension rather than a one-off patch.

Guides packaging DuckDB C extensions with R bindings, including repository layout, installed artifacts, SQL wrappers, generated docs, and native-versus-R responsibility boundaries. Use when building or restructuring an R package around a DuckDB extension.

Guides testing of DuckDB C extensions across SQL, native, wrapper, and external-client layers, with emphasis on real execution paths, sqllogictest coverage, and protocol/interop validation. Use when building trustworthy tests for extensions that expose SQL plus native or service behavior.

Guides vendoring of native dependencies into DuckDB C extensions, including pinning, patch ledgers, static linking, hidden symbols, and compatibility-shim boundaries. Use when an extension needs bundled third-party C/C++ libraries rather than relying only on system dependencies.

Guides exact-compatible rewrites and ports of existing genomics tools into DuckHTS, with pinned upstream validation, attribution, phased scope, and rewrites.bio-style discipline. Use when implementing or maintaining compatibility layers such as mosdepth-, bcftools-, or WisecondorX-aligned behavior in duckhts.

Guides wasm, webR, and duckdb-wasm debugging for DuckHTS, including artifact verification, symbol/export checks, browser-runtime constraints, and package-vs-runtime distinctions. Use when debugging DuckHTS in webR, browser wasm, or duckdb-wasm environments.

Guides design of a DuckDB-native, SQL-first sequencing QC system inspired by RustQC and related upstream tools, with one-pass analytics, reusable kernels, compatibility outputs, and careful threading and summary design. Use when planning DuckQC-style functionality.

Guides design of a DuckDB-native variant effect prediction system, including consequence prediction kernels, transcript/annotation caches, haplotype-aware consequence paths, bcftools csq alignment opportunities, and careful planning for indexes and metadata stores. Use when exploring DuckVEP-style functionality.

Guides genomics tool rewrites and ports that target SQL-native, DuckDB-centered execution using reusable native kernels, streaming readers, parallel scans, alternative file formats, and indexed metadata access. Use when designing genomics capabilities as DuckDB extensions or SQL-first libraries.

Guides AI-assisted bioinformatics rewrites with a library-first, low-dependency mindset, emphasizing battle-tested libraries, C and FFI reuse, innovative composition, portable deployment, and single-pass statistics. Use when designing rewrites that should avoid dependency bloat and maximize reuse.

Guides development of R code and packages using S7 classes, generics, methods, validators, properties, compatibility layers, and package integration. Use when designing or maintaining S7-based APIs or migrating S3/S4 code toward S7.