Claude Skills Library
A library of custom Claude skills — design systems, code audits, documentation pipelines, landing page generators, and more. Each post documents a workstream: what it does, when to use it, and how it works. — scsiwyg.com/claude-skills-and-pipelines-by-worksona.zip
claude-skills
David Olsson
13. From two SKILL.md files to a live landing page in one session
A real session walkthrough: how /landing-page and /deploy turned a bare skills repo into a published site at worksona-repo-publisher-toolkit.netlify.app — in one Claude Code conversation.
12. repo-publisher + scsiwyg: the git-to-blog pipeline
Two skills with a clean division of labour: scsiwyg owns the platform interface, repo-publisher owns the editorial intelligence. Together they turn a git repo into a self-publishing blog.
/repo-publisher
Skill reference: /repo-publisher. Stateful blog orchestrator — reads your git history, scores story candidates, writes drafts, publishes through scsiwyg. Maintains .blog-state/ between runs.
/scsiwyg
Skill reference: /scsiwyg. Platform skill for scsiwyg.com — a headless blog with no web editor. 30 MCP tools for publishing, newsletters, image generation, knowledge graph, and Confluence.
Claude Skills Library
75 reports across 8 pipelines, plus a pair of publishing skills that turn your git history into a live blog. Custom Claude skills for auditing code, documenting projects, analysing competitive moats, reviewing security, measuring UX, modeling infrastructure costs, generating landing pages, and more. Browse by use case or download the full toolkit.
01. How to audit your codebase before it audits you
Your codebase has problems you haven't found yet. Five specialist auditors examine consistency, repetition, security, patterns, and auditability — then grade every finding and hand you a prioritized TODO. Run it again after fixes and get a delta report showing what improved.
/code-audit
Pipeline reference: /code-audit. 9 reports across 5 audit dimensions — consistency, repetition, security, patterns, auditability. Graded remediation plan with delta tracking.
02. What a real security audit looks like
Not a linter. Six parallel agents tear through authentication flows, API endpoints, infrastructure config, protocol-specific attack surfaces, abuse vectors, and OWASP compliance. Produces a posture score and a remediation plan that developers can work through top to bottom.
/security-audit
Pipeline reference: /security-audit. 8 reports from 6 parallel agents — authentication, API/data, infrastructure, protocol-specific, abuse prevention, OWASP compliance. Posture score and remediation plan.
03. Documenting projects that don't document themselves
Twelve numbered reports from a single scan: project overview, tech spec, business benefits, innovation themes, features, extensibility, work zones, technical readiness, and principle alignment. Run it across a portfolio and get cross-cutting synthesis. Documentation that writes itself from what the code actually does.
/doc-suite-generator
Pipeline reference: /doc-suite-generator. 13 reports per project + 7 portfolio synthesis reports. Project overview, tech spec, business benefits, innovation themes, features, extensibility, work zones, readiness, and principle alignment.
04. How defensible is your product, really?
Eight analysts examine your moat from every angle — what's truly differentiated, where competitors will attack, which market inflection points could flip your position, and whether you control the value chain or just rent space on someone else's. Not a SWOT. A structural forensic analysis.
/moat-audit
Pipeline reference: /moat-audit. 9 reports from 8 specialist analysts — moat definition, differentiators, competitive landscape, pressures, innovation imperatives, inflection points, value planes, control planes.
05. Measuring whether your UX actually works
Seven analyzers score your experience across workflow completeness, persona alignment, content coherence, onboarding friction, and code-to-experience gaps. The deliverable is a prioritized improvement plan grounded in the Three-Step Principle: any user, zero to productive, three steps or fewer.
/ux-content-audit
Pipeline reference: /ux-content-audit. 11 reports from 7 analyzers — workflows, personas, onboarding, content, code-to-UX bridge, analytics, headless UX. Experience Readiness scorecard with delta tracking.
06. UX and positioning: is your product saying the right thing?
Five agents audit your content, copy, market positioning, UX patterns, and conversion funnels. Plus SEO. The output is a synthesis action plan sorted by impact × effort — quick wins first, strategic improvements sequenced, and a clear "don't break this" list.
/ux-audit
Pipeline reference: /ux-audit. 7 reports from 5 agents + SEO — content/copy, positioning, UX design, workflows/conversion, synthesis action plan. Impact × effort prioritization.
07. How much does your infrastructure actually cost at scale?
Service topology, per-service cost projections from MVP to 10K users, LOC-to-cost correlation, technology roadmap, and a vendor-agnostic requirements spec. Know where the scaling cliffs are before you hit them.
/infrastructure-cost-audit
Pipeline reference: /infrastructure-cost-audit. 8 reports — service topology, cost projections MVP to 10K users, LOC costing, technology roadmap, vendor-agnostic requirements, recommended stack.
08. Building a landing page from a codebase
Seven agents — app auditor, screenshot analyst, strategy architect, design director, copywriter, page builder, growth advisor — produce a standalone HTML landing page plus the strategy docs, copy deck, rationale, and promotion plan behind it. From codebase to published page.
/landing-page
Pipeline reference: /landing-page. 10 outputs from 7 agents — app audit, screenshot analysis, marketing strategy, design brief, copy deck, standalone HTML page, promotion plan, enhancements, course corrections.
09. Design systems that survive AI generation
Claude builds frontends. Some of them drift — a pure grey sneaks in, a flat black shadow appears, Inter replaces a distinctive serif. Without a system to catch it, every build is a coin flip. Here's how to make the coin flip go away.
/frontend-design-review
Skill reference: /frontend-design-review. Audits frontends against a codified design system across 5 weighted dimensions. Scoring rubric, output structure, trigger phrases, integration points.
10. Tracking SR&ED hours from your git history
Score every git commit against SR&ED qualification criteria, estimate qualifying hours, and produce a CSV ledger ready for your accountant. The git history already has the data — this skill extracts it.
/commit-ledger
Skill reference: /commit-ledger. Scores git commits against SR&ED criteria, estimates qualifying hours, produces CSV ledger for tax credit claims.
11. Shipping to Vercel or Netlify in one command
One command. Detects the project type, picks the right configuration, deploys to Vercel or Netlify, and hands you a live URL. Static sites, Next.js, SPAs. Build verification before deploy.
/deploy
Skill reference: /deploy. Detects project type, verifies build locally, deploys to Vercel or Netlify. Two sub-skills with platform-specific guides.
subscribe