Agent
Claude Code
Skills, commands, and rules designed for Claude Code — Anthropic's CLI coding agent. Install via CLAUDE.md or the Cendis CLI.
20 skills available.
Distinctive Frontend Design
Anthropic's official skill for generating production-grade frontend interfaces with bold aesthetic direction — avoids generic 'AI slop' UIs in favor of intentional, distinctive design.
Ban mock.module in Bun Tests
Stops Claude Code from reaching for Bun's mock.module — which poisons the global module registry and breaks test isolation. Forces factory + dependency injection patterns instead.
Full Site SEO Audit
Runs a complete SEO audit by dispatching parallel subagents across 10 specialist areas — technical, content, schema, GEO, performance, local — and produces a prioritized action plan with a 0-100 health score.
Competitor / Alternatives Pages Generator
Generates 'X vs Y' and 'alternatives to X' SEO pages that capture high-intent comparison traffic — with honest framing, structured comparison tables, and conversion CTAs.
Content E-E-A-T Review
Reviews content for E-E-A-T signals (Experience, Expertise, Authoritativeness, Trust), readability, depth, and AI citation readiness — based on Google's 2025 Quality Rater Guidelines.
GEO / AI Overviews Optimization
Optimizes content and site infrastructure to be cited by Google AI Overviews, ChatGPT, Perplexity, and Bing Copilot — passage-level citability, llms.txt, brand mention signals, and crawler accessibility.
Programmatic SEO Planner
Plans and scaffolds programmatic SEO at scale — template + data → hundreds or thousands of unique landing pages — with built-in thin-content guards so you don't poison your domain.
Schema.org Structured Data
Detects, validates, and generates JSON-LD structured data for any page — Article, Product, Organization, FAQ, HowTo, BreadcrumbList, and more — to unlock rich results and AI citations.
Technical SEO Audit
Audits the technical layer of SEO — crawlability, indexability, canonical tags, sitemaps, security headers, mobile rendering, JavaScript-rendered content, and Core Web Vitals.
Ship PR Pipeline
Mandatory pre-PR pipeline for Claude Code — branch check, simplify, security review, tests, scope check. Blocks 'merge-ready' status until every gate is green.
Story-Driven Development
Requires Claude Code to find a story with acceptance criteria before writing any feature code — and to convert each AC into a failing test before implementation.
Subagent Model Strategy
Routes Claude Code subagent work to the right model — Opus for architecture and security review, Sonnet for grunt work like exploration, search, and boilerplate. Cuts cost ~5x with no quality loss.
Strict TDD Story Loop
Forces Claude Code into strict red/green TDD — write failing tests first, commit red, then implement to green. No retrofitting tests after the fact.
Verify Before Done
Forces Claude Code to run typecheck and the full test suite before marking any task complete. Eliminates 'it should work' hand-waving.
Verify Environment Before Debugging
Stops Claude Code from misdiagnosing production bugs as local-dev issues — confirms which environment the report came from before proposing a fix.
Error Escalation (3-Strike Rule)
Stops Claude Code from spinning on errors — after 3 varied attempts at a problem, the agent surfaces the issue in plain language instead of burning tokens.
Fresh Start Session Handoff
Generates a self-contained handoff prompt at the end of every session so a new Claude Code conversation can pick up exactly where the last one left off.
Plan Before Acting
Forces Claude Code into plan mode for any non-trivial task — write specs upfront, reduce ambiguity, and catch architectural issues before touching code.
PR Handoff Prompt
At the end of every PR cycle, Claude Code generates a complete handoff prompt so the next conversation can pick up review feedback, CI results, or follow-up work without re-discovery.
Scope Discipline
Prevents Claude Code from expanding scope mid-task — no drive-by refactors, no unrelated fixes, no editorial changes cascading into renames.