wdm0006/python-skills

Plugin marketplace

Opinionated, checklist-driven skills for professional software development across languages - Python libraries and web apps, Go projects, Swift/Apple apps, Rust crates, Scala projects, browser extensions, and UI design

GitHub stars: 111
plugins: 28
last updated: Nov 11, 2026

View on GitHub

Add marketplace

/plugin marketplace add wdm0006/python-skills

Run it inside Claude Code, then install plugins from it with /plugin.

Plugins

1browser-extensionsBuild and test Manifest V3 browser extensions - service-worker message channels that survive teardown, lost-update-safe chrome.storage writes, settings validation in the popup, a Jest chrome mock that can actually fail, plus git hygiene

2calibrating-thresholds-and-baselinesCheck that a score-threshold-flag pipeline separates anything before trusting or tuning it - reference constants that are hand-authored but formatted as measurements, a feature whose units do not match the constant it is compared against, per-class fire rates exposing a flag that fires on nearly every item of both classes, rate features confounded by item length, confidence scores where one feature family contributes once per sub-feature, threshold sweeps hardcoded to one comparison direction and one statistic, and separating decision-carrying features from display-only ones

3deferring-heavy-dependenciesKeep expensive resources off code paths that never use them - constructor work every caller pays for, one-size factories that load the heaviest member for the cheapest request, load-and-unload-per-call caches that are not caches, per-call timing in a fresh process, telling a one-time import from a repeated model load, and the test fallout of making a field lazy

4go-projectsSet up and maintain Go projects with a CI gate that actually gates - module-path correctness, golangci-lint config matched to the installed major, deterministic gofmt, a pinned toolchain, meaningful test jobs, plus git hygiene

5guarding-destructive-operationsPut real preconditions on operations that delete, overwrite, or rewrite history, and on code that turns a caller-supplied name into a path - refusing instead of warning, guards placed ahead of the first mutation, structural ownership checks rather than string prefixes, name validation plus resolved-path containment as two independent checks, the deliberate dry-run gap, and mutation-testing each half separately

6merging-concurrent-branchesResolve conflicts when several branches are open against one repo at once - the hotspot files every change touches, additive registries unioned instead of chosen, version counters recomputed from the integration branch, generated artifacts rebuilt rather than merged, deterministic serialization so diffs stay reviewable, shared identifiers appended and never renumbered, and re-running the real gate because a clean auto-merge is not a passing test

7normalizing-text-for-measurementStrip, parse, and segment a document before something else measures it - close tokens that emit no separator and fuse two blocks into one word, character-based metrics inflated by leftover syntax while word-based metrics break on boundaries the strip never inserted, sibling consumers that skip the normalizer, parser presets whose rule set reclassifies a construct, sections keyed by a non-unique heading that collapse and double-count into an ancestor, deliberate lossiness mistaken for a bug, and regression tests that assert markdown-equals-plain-equivalent instead of a float from an unpinned library

8permissioning-ci-workflowsGive a CI job exactly the credentials it needs and know when it has none - explicit permissions blocks instead of the repo-wide default token scope, the 403 header that names the missing scope, one API call needing two scopes, fork pull requests running with a read-only token and no secrets, event-conditioned steps that skip on the re-run and go green, why pull_request_target is not the workaround, and preferring checks that need no credential

9python-library-completeComplete Python library development toolkit - all Python skills for building professional-grade libraries, web apps, and MCP servers

10python-library-distributionSkills for packaging, releasing, and distributing Python libraries

11python-library-foundationsCore skills for starting Python library projects - setup, quality, and testing

12python-library-qualityQuality-focused skills - security, performance, API design, and git hygiene

13python-llm-featuresBuild application features on top of an LLM API - prompt surfaces that drift, filtered evaluator sets that cannot silently become clean passes, model config wiring, live-model tests kept out of CI, stochastic accuracy evaluation, and truthful presentation of model output

14python-mcp-serversBuild robust Python MCP servers - tool error contracts, event-loop-safe blocking work, subprocess wrapping, single-file vs packaged distribution, and global-state-free testing

15python-web-appOpinionated reference architecture for building production Python web apps - FastAPI, async SQLAlchemy, Stripe, and Dockerized deployment

16reporting-derived-metricsCompute and report statistics, scores and flags from samples that may be too small - or too heavy-tailed - to support them: undefined dispersion returned as 0.0 and tripping a minimum threshold, sentinel choice, threshold blocks gated on whether the value was measured, reports that narrate findings from absent data, nullability as a public API change, broad excepts that disguise a metric bug as a normal result, and cohort comparisons where one outlier makes the mean point the opposite way to the median

17reproducing-ci-locallyMake the local check agree with the runner - deriving the command, paths, markers and env from the workflow file rather than the Makefile, unblocking gate steps that short-circuit and hide the next failure, pinning the linter version CI resolves, building the interpreter and extras the runner builds, and confirming a run is green instead of explaining a red job away

18resumable-sync-jobsWrite batch and sync jobs that fail honestly - exit codes that report partial failure, checkpoints that are safe to resume, per-item tolerance vs fatal abort, unknown-vs-zero in aggregated output, bounded retries, honest dry-runs, and compensating reserved quota

19rust-cratesSet up, test, and publish Rust crates - Cargo layout, a fmt/clippy(-D warnings)/test CI gate, MSRV pinning, feature-flag hygiene, crates.io publishing, plus git hygiene

20scala-projectsSet up, test, and publish Scala projects - sbt layout, scalafmt/scalafix gates, Scala 2-vs-3 cross-building, real test jobs, Maven Central publishing via sbt-ci-release, plus git hygiene

21shipping-across-surfacesLand a change everywhere the same fact is stated - the full surface inventory from landing copy to untyped frontend consumers of typed responses, generating a surface instead of restating it, drift tests against the live registry, never hand-maintaining a copy of a surface another codebase owns, paired PRs for format changes, and keeping overloaded product words apart

22shipping-build-artifactsMake the build step a real gate on what you distribute - scripts that fail instead of warning on a missing input, size checks bounded on both sides, file lists checked against the entrypoints they must cover, committed bundles that can't go stale, portable release shell, and artifact verification that runs before the upload

23shipping-swift-appsShip Swift apps to TestFlight and the App Store - one App Store Connect API key for fastlane and xcodebuild, metadata-only lanes that cannot submit a build, headless archive and upload, build numbers read from the API instead of guessed, closed pre-release trains, SKIP_INSTALL for bundled helper apps, and per-destination archives for multiplatform targets

24swift-appsBuild, test, and ship native Swift/SwiftUI apps for Apple platforms - unsigned CI builds, the hand-maintained pbxproj, SwiftPM-vs-app-target test boundaries, Keychain/OAuth security, CloudKit/SwiftData constraints, TestFlight releases with fastlane and xcodebuild, plus git hygiene

25ui-designMake products look and feel finished - design tokens and type/spacing/color scales before restyling, a modernization playbook for dated UIs, accessible controls with measured contrast, designed empty/loading/error states and microcopy, and landing/pricing pages that explain the product fast and convert honestly

26verifying-external-behaviorConfirm what a third-party library, remote API, build backend, or scraped document actually does before depending on it - throwaway probes, permissive clients that forward wrong arguments, per-endpoint docs that don't generalize, fakes that encode your assumption, and dry-runs that skip the failing step

27wiring-application-configLoad, merge and validate layered application configuration so a mis-wired or misspelled key fails loudly instead of reverting to the default - sections the code reads that exist nowhere, sections nothing reads, unvalidated recursive merges, bool passing as int, defaults aliased by a shallow merge, falsy-but-valid values killed by or/||, partial write-backs stickier than a missing file, and bootstrap ordering that swallows the loader's own warnings

28writing-defect-reportsEstablish a finding before you publish it and correct it after - headlines that overstate what reproduces at the layer a user sees, reporting code no entry point reaches, guards that only match the fixture written beside them, caveats the project's own records already answered, prior notes re-verified against the tree rather than against the note, the narrowest injection point that reproduces a failure, probe output a harness swallows, red gates described as pre-existing, and withdrawing a published claim in the thread where it was published

Details

Repository: github.com/wdm0006/python-skills
Categories: community
Listed since: Feb 24, 2026