Josh Hornby
Community@joshhornby · Manchester, UK
Josh Hornby publishes 43 agent skills for TypeScript architecture, behavior-driven testing, spec-driven planning, and technical writing quality.
Agent Skills by Josh Hornby
Showing 42 vetted skills indexed across 1 GitHub repositories.
joshhornbyjosh-writing-voice
Rewrites prose in Josh's personal writing voice with plain language, active voice, and British spelling.
joshhornbytechnical-writing
Writes and reviews developer-facing documentation using reader-first structure and verifiable claims.
joshhornbypanel-review
Reviews code diffs by fanning out sub-agents that each apply one installed skill as a review lens.
joshhornbytypescript-strict
Enforces strict TypeScript patterns using discriminated unions, branded types, and schema-first boundary validation.
joshhornbyclaude-doctor
Analyzes recent session learnings and creates a PR proposing Claude configuration improvements.
joshhornbyfront-end-testing
Write behavior-driven UI tests using Vitest Browser Mode, Playwright E2E, and Testing Library.
joshhornbystoryboard
Generates a single HTML page embedding related UX mocks with audit checklists, flow diagrams, and gap cards.
joshhornbyubiquitous-language
Enforces one glossary-driven ubiquitous language per bounded context through a five-step decision protocol.
joshhornbyevaluate-existing-solutions
Compares libraries, services, and platform primitives against a bespoke baseline for build-versus-adopt decisions.
joshhornbycharacterisation-tests
Write tests that document the actual current behavior of untested legacy code before refactoring.
joshhornbydiagrams
Create diagrams and visualizations in Markdown using Mermaid, Graphviz, Vega-Lite, PlantUML, and HTML templates.
joshhornbyhexagonal-architecture
Implements TypeScript ports and adapters with dependency inversion and domain isolation.
joshhornbyfinding-seams
Identifies substitution points in tightly-coupled TypeScript code to make legacy dependencies testable.
joshhornbydouble-check
Obtains an independent read-only second-opinion review of finished work with mandatory scope-fidelity checks.
joshhornbyimprove-codebase-architecture
Audit a repository for evidence-backed architecture improvements and present ranked candidates in a self-contained HTML report.
joshhornbyplanning
Plans implementation work as vertical slices with TDD increments and small reviewable pull requests.
joshhornbyexpectations
Capture learnings, gotchas, and architectural decisions into the correct project documentation files.
joshhornbyfind-skills
Discover, inspect, and install agent skills from the open skills ecosystem.
joshhornbycodebase-design
Design deep modules with small, stable caller-facing contracts and justified seams.
joshhornbyprototype
Build throwaway terminal or UI prototypes to validate logic and design questions.
joshhornbystack-pull-requests
Plans and manages stacked pull requests with dependent branch lineage on GitHub.
joshhornbyapi-design
Designs stable REST API contracts with RFC 9457 errors, idempotency, and versioning guidance.
joshhornbyspecification
Facilitates one-question-at-a-time specification conversations that produce example maps and acceptance criteria.
joshhornbyto-issues
Converts plans and PRDs into vertical-slice issues on a project issue tracker.
Frequently Asked Questions About Josh Hornby
FAQPage SchemaWhat tasks can I accomplish with Josh Hornby's skills?▼
You can design strict TypeScript domain models, audit and restructure codebases, run behavior-driven and mutation testing, split epics into vertical slices, plan stacked pull requests, review PRs through multi-agent panels, and edit prose to remove AI-writing patterns.
Who are these skills designed for?▼
Software engineers and technical leads working in TypeScript codebases who practice TDD, DDD, or hexagonal architecture, plus developers writing READMEs, guides, and release notes who want skimmable, evidence-backed documentation.
How do the review and verification skills work in practice?▼
Invoke /panel-review on a branch, diff, or PR; an orchestrator fans out sub-agents each loading one skill as a review lens, adversarially verifies findings, and returns one ranked report. The double-check skill adds a cross-provider second opinion with a scope-fidelity check.
Are these skills free and open source?▼
Yes. The skills are published publicly on GitHub under the joshhornby account, and the diagrams skill explicitly notes MIT-licensed origins adapted from markdown-viewer. No paid tier or license fee is indicated in the manifest.
What prerequisites do these skills require?▼
They run inside a Claude Code-style agent environment with Read, Write, Edit, Grep, Glob, and Bash permissions. Specific skills need project context: Stryker for mutation-testing, Vitest Browser Mode and Playwright for front-end-testing, and an issue tracker for to-issues and to-prd.