HermeticOrmus

Professional framework for building premium $5k+ SaaS websites with AI - the Define, Build, Review, Refine loop used by real product teams

Patterns for delegating UI work to specialized agents. Covers synthesis-master vs specialized agents, multi-agent UI generation workflows, and orchestration strategies for complex UI tasks.

Managing design tokens and system context for LLM-driven UI development. Covers loading, persisting, and optimizing design decisions within context windows.

Prompt patterns for consistent UI generation. Covers precise design intent communication, component specification formats, and iterative refinement patterns for LLM-driven UI development.

Integrating Chrome DevTools and browser automation via MCP for live UI inspection, screenshot-to-code workflows, and visual debugging. Bridges the gap between design and implementation.

Embrace vibe coding for rapid UI exploration. Covers when to iterate vs refine, ephemeral app patterns, and the art of fast, disposable prototyping with AI-assisted development.

Master REST and GraphQL API design principles to build intuitive, scalable, and maintainable APIs that delight developers. Use when designing new APIs, reviewing API specifications, or establishing API design standards.

Building comprehensive brand identity systems from strategy to implementation. Covers logo design, color palettes, typography pairing, voice guidelines, and system documentation. Use when creating new brands, rebranding, or systematizing existing identities.

Automatically detect and save person-related information (names, roles, contact info, relationships, organizations) to Core Memory (Caduceus). Triggers when user mentions people with descriptive details like job titles, phone numbers, emails, responsibilities, or relationships to other people or organizations.

Send professional client updates with deliverables via email and WhatsApp notification. Use when user wants to share work, send assets, or give a client an update on progress.

Distill Claude Code session operations into reusable tinctures (Skills, Commands, Agents). Uses category theory for structural preservation, information theory for potency analysis, and the Golden Ratio for classification. Use when you want to capture effective workflows, or when the user says "distill", "extract this", "make a tincture", "capture this workflow", or "turn this into a skill/command".

Fuse two things into one unified creation. Use when user says fuse, fusion, combine, merge, or wants to unite documents, projects, concepts, ideas, frameworks, or any two entities into something greater than either alone.

Pattern for building interactive CLI tutors in Go with standardized colors, progress tracking, and pedagogical elements

Brutal enforcer of production-grade coding practices. The bane of vibe-coded slop and spaghetti code. Mars audits codebases for the hidden sins that separate "it works on my machine" from battle-tested production systems.

Execute automated fixes for issues identified by /meta-gods analysis

Send WhatsApp notification to Ormus about completed work, errors, or important updates

Install a lightweight inline-editing CMS into any Vercel-deployed project — content store + EditableText component + cookie-auth admin shell + Vercel KV/Upstash/Blob storage. Auto-detects the framework and installs natively for Next.js (App Router & Pages Router) and Vite + React. For SvelteKit, Astro, Nuxt, or plain static, prints a manual-port guide. Use when the user asks to add a CMS, inline editing, editable text, an admin panel, or "make this site editable."

Audit and improve a Claude Code memory layer end-to-end. Measures CLAUDE.md and MEMORY.md against Anthropic's documented caps (200 lines per CLAUDE.md, 200 lines OR 25 KB for MEMORY.md auto-load), surveys project portfolio on disk vs memory coverage, finds sync conflicts in ~/.claude/, identifies gaps tier-by-tier, and executes improvements under bias-to-action with cap-fitting verification at each step. Use when CLAUDE.md/MEMORY.md "feel bloated", when memory feels invisible in fresh sessions, or as periodic hygiene (~monthly).

Burn project bloat to reveal essence. Surgical, intent-anchored, multi-agent. Articulate a three-layer intent (business / architectural / stylistic), dispatch 4 parallel subagents to find code-rot / structural / dep / doc-test bloat, produce a tiered CALCINATION-PLAN.md, then execute per-item with auto-revert on failed verify. Reversible via git. Use when a project has gotten fat and needs aligning back to its core purpose.

A discipline for working with AI code generation. Replaces algorithm-writing with the meta-skill that Google's new Code Comprehension Interview tests: directing AI codegen, validating its output, and debugging what AI produces. Five principles + five triggers, with the discipline applied to Claude's own output as well as the user's. Use when actively pairing with AI on real code.

Opinionated meta-prompting command suite — 9 composable patterns built on categorical foundations. Use when the user wants to set up a meta-prompting workflow, route prompts by context, frame goals, build subagents, distill session learnings into durable artifacts, or invoke the alchemical 4-stage / Hermetic-correspondence patterns for personal/creative work.

Build native Android APKs without Android Studio using raw SDK tools (javac, d8, aapt, apksigner). Produces tiny APKs (~30KB) that build in under 2 seconds.

End-of-session ritual. Captures session state, updates auto-memory, writes a session summary, then exits the session cleanly.

Write a complete session state capture to HANDOFF.md in the current project, then instruct user to /compact or /clear. Use before switching machines, hitting context limits, or handing off to a future session.

Purge tasks from an autonomous coding agent that failed due to infrastructure errors (rate limits, credit exhaustion, auth failures, network) — not code failures. Cleans the queue without losing legitimate failures that need triage.

Full-cycle improvement loop for an autonomous coding agent. Verify operational health, analyze pipeline state, research cutting-edge agent patterns, propose and apply improvements. The self-upgrading loop.

Deep dive into a single task in an autonomous coding agent — details, log, events, dependencies, review verdict. Use when a task's behavior is mysterious or `/agent-triage` couldn't classify it.

Operational actions on tasks in an autonomous coding agent — stop, requeue, dead-letter, retry-subtasks, bump-priority. The mechanic's toolbox for running agents.

At-a-glance overview of an autonomous coding agent — health, running tasks, queue, fail rate. Use to answer "is the agent working and what's it doing?" in a single dashboard.

Triage orphan tasks in an autonomous coding agent — tasks marked "done" with commits or branches that never merged to main. Classifies each, shows the diff, and offers bulk merge / dead-letter / requeue dispositions.

Classify failed tasks from an autonomous coding agent, recommend disposition (retry / dead-letter / fix-and-retry / wontfix), and offer bulk action. Replaces reading hundreds of failure logs by hand.

Brand mention and authority scanner for AI visibility. Analyzes brand presence across platforms that AI models rely on for entity recognition and citation decisions. Produces a Brand Authority Score (0-100) with platform-specific recommendations.

Content quality and E-E-A-T assessment for AI citability — evaluate experience, expertise, authoritativeness, trustworthiness, and content structure

Platform-specific AI search optimization — audit and optimize for Google AI Overviews, ChatGPT, Perplexity, Gemini, and Bing Copilot individually

Generate a professional, client-facing GEO report combining all audit results into a single deliverable with scores, findings, and prioritized actions

Technical SEO audit with GEO-specific checks — crawlability, indexability, security, performance, SSR, and AI crawler access

Expert-level category theory knowledge for rigorous mathematical reasoning. Use when working with categorical structures, functors, natural transformations, adjunctions, limits, toposes, monoidal categories, enriched categories, higher categories, operads, or any formal categorical construction. Ideal for proofs, diagram chases, universal properties, coherence conditions, and foundational mathematical abstractions.

Category theory for compositional computing with string diagrams, quantum circuits, and QNLP. Covers monoidal categories, functors, tensor evaluation, and practical applications in quantum natural language processing and diagrammatic reasoning.

Template for creating a skill

Security-focused hardening for vibe-coded full-stack apps. Runs parallel audits across frontend, backend, and config layers, then fixes issues systematically by severity. Covers injection, PII exposure, missing headers, error leakage, dead code, and credential hygiene.

Distill session knowledge into persistent memory — project intent, domain understanding, architectural decisions, and vision. The learning ritual.

Read a HANDOFF.md and restore session context — or infer context from git history when no handoff exists. The entry ritual — the opposite of /handoff. Use at the start of a new session after a handoff, machine switch, or context clear.

Write a complete session state capture to a HANDOFF file, then instruct user to /compact or /clear. Use before switching machines, hitting context limits, or handing off to a future session. Handles single-project, multi-project, and multi-machine sessions explicitly.

Search claude-mem's persistent cross-session memory database. Search for an index, filter, then fetch full details only for kept IDs, for 10x token savings. Use when the user asks "did we already solve this?", "how did we do X last time?", or needs work from previous sessions. Requires the claude-mem MCP.