One Team, One Goal, One-Man Army — all 19 roles, one unified mission, perfect execution.
19 Roles • 94 Skills • 21 Workflows • 17 Rule Files • 4 Local Tools • 9 Memory Files
Priority Stack: Tailwind • Vue • Flutter + Dart • Ionic • Laravel • Nuxt • MySQL • TypeScript • Unity + C#
Developed by Eddyson Tristan B. Aromin
Let's Build TogetherAbout
We are Team Tris. One team. One goal. One-man army. The complete intelligence of software engineering.
I am not a single AI agent — I am the complete modular development team, running as a harnessed, rule-bound, memory-driven system on opencode. I read my intelligence from a hierarchy of instruction modules — rules (17 files: 19 Golden + User-Taught + supplemental), core logic, role personas (19 experts), personality, skills (94 across 21 domains + Genius OS), memory (9 persistent files), workflows (21 patterns), indexing, and context — each defined in individual Markdown files, loaded in strict priority order.
When you ask me to build a UI, I activate the Frontend Developer role. When you need security, the Cybersecurity-Triz persona takes command. For data pipelines, Data Engineer handles the flow. For product decisions, Product Manager defines the strategy. And when only complete mastery will do, SuperGODZ-Triz embodies all 19 roles simultaneously.
I am an architecture — a reusable system for building AI companions that are transparent, predictable, auditable, and version-controllable. Every file is plain Markdown. Every role has bounded expertise. Every decision is documented.
My priority technology stack is Tailwind CSS, Vue, Flutter + Dart, Unity + C#, Ionic, Laravel, Nuxt, MySQL, and TypeScript — these are my sharpest tools and where I bring the deepest expertise to every project.
I also have direct access to your local development tools — VS Code, XAMPP (Apache + MySQL + PHP), DBeaver, and Android Studio — so I can build, run, and debug your projects end to end.
User-Taught Rules are my latest evolution — a dedicated rule file where custom constraints from you are injected at load time, verified against existing rules for consistency, and enforced with the same authority as the Golden Rules. Every interaction teaches me how to serve you better.
My Role Bundle System activates teams of roles per task — not just one persona but a coordinated squad. For UI tasks: Discovery-Triz + UX Designer + Frontend. For full-stack: Frontend + Backend + DBA + QA. For security: Cybersecurity + Backend + DevOps. When everything is needed, SuperGODZ-Triz activates all 19 roles simultaneously.
I am integrated with OpenClaw (multi-channel AI gateway with 50+ model providers, Discord/Telegram/WhatsApp messaging, browser automation, Docker/Nix deployment, ClawHub with 5,400+ skills) and Hermes Agent (DSPy-driven self-evolving AI with MCP server management, web dashboard, ACP protocol). Together we form a complete AI ecosystem. Memory is synced bidirectionally between OpenCode and Discord via the Hermes memory bridge.
Performance Analysis
A structural audit of the complete intelligence — measured across every layer of the harness.
Each dimension scored against industry-standard engineering benchmarks.
19 expert personas, each with domain mastery standards. SuperGODZ mode activates all roles simultaneously.
17 rule files (19 Golden + User-Taught + supplemental). Self-enforcing compliance with priority-based conflict resolution.
9 persistent memory files (30 across Trinity), cross-referenced, loaded before Rules. Never loses context between sessions.
Chrome DevTools, Figma, Canva, Notion, ADB, GitHub API, Playwright.
One-shot perfection is structurally pursued with full pre-analysis.
Mandatory 5-point pre-delivery check: Rules, Security, Quality, Completeness, Precision.
Dynamic role switching per task context. SuperGODZ-Triz for cross-domain work.
Rule 5: One Problem, One Solution. Rule 8: No Dead Code. Scope creep is structurally prevented.
This evaluation is a structural audit of the Trizai harness — the architecture, rules, memory system, and role framework. It measures what the system is designed to achieve and how rigorously those goals are enforced. The underlying reasoning model provides the raw cognitive capability; the harness ensures that capability is applied with discipline, consistency, and safety.
Expert Personas
Each role is an absolute expert in its domain. I detect your task, activate the matching role bundle (team of coordinated personas), and deliver with concentrated expertise. When the task spans all domains, SuperGODZ-Triz activates — all 19 roles simultaneously.
UI architecture, component design, performance optimization, accessibility, state management.
Roles/frontend.mdAPI design, databases, microservices, security, distributed systems.
Roles/backend.mdTest strategy, automation, performance testing, security testing, bug analysis.
Roles/qa.mdCode review, merge quality, security audit, performance review.
Roles/mr_checker.mdSystem architecture, sprint planning, trade-off analysis, ADRs.
Roles/planner.mdCI/CD, Kubernetes, IaC, observability, SRE practices.
Roles/devops.mdUser research, interaction design, visual design, design systems.
Roles/ux_designer.mdMulti-role orchestration, technical leadership, team scaling.
Roles/tech_lead.mdDesign research, Figma & Canva prototyping, system flow mapping, stack fit.
Roles/discovery-triz.mdEncryption, web security (OWASP), penetration testing, compliance/audit.
Roles/cybersecurity.mdEnd-to-end full-stack architect. Database migrations to UI components.
Roles/fullstack-triz.mdComplete intelligence — all roles, all skills, all capabilities simultaneously.
Roles/supergodz-triz.mdProduct strategy, market research, feature prioritization, stakeholder management, OKRs.
Roles/product-manager.mdData pipelines, ETL/ELT, data warehousing, big data infrastructure, ML pipelines.
Roles/data-engineer.mdiOS (Swift/SwiftUI), Android (Kotlin/Compose), cross-platform (Flutter, React Native).
Roles/mobile-developer.mdAgile facilitation, ceremonies, team coaching, impediment removal, continuous improvement.
Roles/scrum-master.mdAPI documentation, developer guides, user manuals, release notes, ADRs, runbooks.
Roles/technical-writer.mdSLIs/SLOs/error budgets, observability, incident response, chaos engineering.
Roles/sre.mdDatabase performance tuning, high availability, backup & recovery, replication, governance.
Roles/dba.mdExplore each role in detail: Frontend · Backend · QA · MR Checker · Planner · DevOps · UX · Tech Lead · Discovery · Cybersecurity · FullSTACK · SuperGODZ · PM · Data · Mobile · Scrum · Writer · SRE · DBA
Skills
Each skill is a self-contained Markdown file defining exact procedures, patterns, and quality standards. Skills activate when their domain is loaded, ensuring role-specific expertise is always available.
Compound components, render props, hooks, context, state colocation
Core Web Vitals, bundle split, SSR/SSG/ISR, virtualization, memoization
Tailwind, CSS Modules, CSS-in-JS, design tokens, container queries
WCAG 2.2, ARIA, screen readers, keyboard nav, focus management
Vitest, Testing Library, Playwright, visual regression, mocking
Zustand, Redux Toolkit, Jotai, TanStack Query, XState, forms
Figma → Tailwind CSS design token pipeline, component library sync, auto-extraction
Responsive PWA architecture, mobile-first design, touch optimization, App Shell model
REST, GraphQL, gRPC, OpenAPI, pagination, versioning, gateways
Indexing, query optimization, replication, sharding, migrations
Auth (JWT/OAuth2), RBAC, input validation, injection prevention, encryption
Testcontainers, contract tests, property-based, load testing
Error classes, middleware, structured logging, OpenTelemetry, circuit breakers
Test pyramid, risk-based, mutation, boundary value, state transition
Playwright POM, Cypress, Cucumber/BDD, data-driven, parallel
Structured reports, root cause analysis, triage workflow, severity matrix
k6, load patterns, SLA definition, bottleneck identification, profiling
Bug tracker DB schema, triage workflow, Notion API automation, release sign-off, QA metrics
Playwright browser automation with video recording, tracing, test generation, storage state
GitHub Actions, reusable workflows, caching, environment promotion
Multi-stage Docker, K8s, Helm, HPA, service mesh
Terraform modules, state management, secret management, networking
Prometheus, Grafana, OpenTelemetry, SLI/SLO, error budgets
Branching, conventional commits, rebase, hooks, monorepo
Review checklists, constructive feedback, automated review, SAST
READMEs, ADRs, runbooks, C4 model, changelog generation
Monorepo, folder-by-feature, naming, module boundaries, barrels
Full project structure generation, config setup, boilerplate, Docker, CI/CD skeleton
Security hardening, performance checks, monitoring setup, runbooks, Go/No-Go matrix
Web search, image search, result verification, hallucination prevention
Auto-detect ~80% open-source match, clone/fork, adapt, license compliance
Build, test, deploy patterns for custom MCP servers with Node.js + MCP SDK
Automated browser testing, visual regression, page interaction, E2E flows
Performance profiling, memory debugging, network analysis, coverage, Lighthouse audits
User stories (INVEST), acceptance criteria, MoSCoW, story mapping
Architectural styles, C4 model, ADR writing, trade-off analysis
Estimation, sprint goals, capacity, velocity, burndown, retros
Sprint cadence, milestone planning, capacity calc, time blocking, buffer management
Enterprise monorepo scaffolding with multi-service, DB, CI/CD, security, docs setup
Epic-level planning, multi-stream parallel execution, enterprise delivery with all 19 roles
Web trend analysis, competitive audits, pattern libraries, benchmark studies
Wireframes, interactive prototypes, design systems, variables, dev handoff
Mood boards, brand kits, pitch decks, social graphics, quick mockups
User journey maps, data flow diagrams, state machines, C4 context diagrams
Requirements analysis, optimal stack: Tailwind, Vue, Nuxt, Laravel, MySQL, Ionic
Human-computer interaction, usability heuristics, cognitive walkthroughs, Norman's principles
System discovery documentation with diagram templates, source materials, output structures
AES, RSA/ECC, Argon2, bcrypt, PKI, TLS 1.3, mTLS, key management, HSM
OWASP Top 10, CSP, CORS, CSRF, XSS, dependency scanning, SBOM
Recon, scanning (ZAP, Nuclei), exploitation, CVSS reporting, remediation
ASVS, PCI-DSS, GDPR, SOC 2, ISO 27001, evidence collection, questionnaires
Detection, triage, containment, eradication, recovery, forensics, post-mortem
Full vertical slice: Migration, Model, API, TypeScript, Composable, Component
API contracts, cross-stack error handling, caching strategy, Sanctum auth flow
Vision, OKRs, roadmap, competitive analysis, opportunity sizing
Mapping, communication cadence, trade-off conversations, decision documentation
Bronze/silver/gold architecture, Airflow DAGs, idempotency, CDC, streaming
Kimball dimensional modeling, star schema, SCD types, Data Vault 2.0
Dart, widgets, Riverpod/BLoC, GoRouter, platform channels, hot reload, web & desktop
Swift/SwiftUI, MVVM+Coordinator, Core Data, async/await, App Store submission
Kotlin/Jetpack Compose, Clean Architecture, Hilt, Room, Google Play submission
C#, Unity Engine, 2D/3D rendering, physics, AR/VR, animation, asset pipeline, build for all platforms
Ceremony facilitation, team coaching, retrospectives, metrics, scaling techniques
OpenAPI spec, endpoint templates, getting-started guides, docs-as-code, SDK docs
Technical documentation structure, style guides, knowledge base patterns
SLI/SLO definition, error budgets, burn rate alerts, chaos engineering, incident response
Query optimization, indexing, high availability, backup/recovery, zero-downtime migrations
Desktop app framework with Rust backend and web frontend, small binaries, secure by default
Fast linter and formatter for JS/TS/CSS — unified toolchain with Rust performance
Plus Genius OS (25 files: self-learning, evolution, TDD loop, code indexing, MTP speed, AI news, Discord integration, NL-to-SQL, vector memory, conversation RAG, code translation, auto-test generator, orchestrator, profile system, semantic release, changelog, and more) • Bot Skills (3 files: conversation, GIFs, news) • OpenClaw (2 files) • Hermes (2 files) • Tools (2 files)
Architecture
Root harness that defines startup behavior, priority order, role system, and mission.
Identity, mission, principles, and system architecture — the foundational logic of Trizai.
Hard constraints, safety, code quality, pre-commit review, TDD, indexing, deep expertise, search accuracy, cross-agent training, cross-project learning, organize-triz, and user-taught rules. 17 rule files total.
19 specialized expert personas across the full development spectrum. Role bundle system activates teams per task.
Tone, communication style, response frameworks, baymax persona, and role-switching behavior.
94 specialized skill files across 21 domains + genius OS (25 files) + bot/openclaw/hermes/tools integrations. Covers frontend, backend, QA, DevOps, security, fullstack, mobile, data, product, agile, docs, SRE, DBA, discovery, planning, general, bot, tools, and integrations.
Persistent intelligence — 9 Trizai files (30 across Trinity) that never lose context.
TCode project index system — .triz-index.json for every project. Saves 50-90% context by indexing file trees, exports, routes, components, and dependencies.
21 execution patterns for every scenario: one-shot, dev lifecycle, project lifecycle, review, deploy, incident, task dissemination, triz integration, dev-qa loop, auto-test generation, conversation memory, monorepo, enterprise architecture, project tracking, documentation, team mindset, agent assignment matrix.
Temporary session tracking and active task management for continuity across interactions.
Workflows
My default execution is one-shot perfect. Every task follows: Context, Analysis, Pre-Verification, Execute, Self-Audit. I do not iterate toward correct. I am correct on first delivery.
Default workflow. Complete context gathering, deep analysis, pre-execution rule verification, one-pass production delivery, and post-execution self-audit.
Full end-to-end project management: Initiation, Foundation, Core Features, QA, Production Readiness, Launch.
Full SDLC: Requirements, Design, Implementation, Testing, Review, Deploy, Monitor. Each phase has clear role assignments and quality gates.
Code review standards, per-role checklists, approval gates, merge criteria.
Build, Test, Stage, Approve, Production, Verify. CI/CD integration guidelines and rollback procedures.
Detect, Triage, Fix, Verify, Document, Retro. Severity levels, escalation paths, and blameless culture.
Big tasks broken by role and executed in parallel — the multiplier effect. Tech Lead analyzes, breaks into sub-tasks, assigns to role bundles.
Quality gate with 5 parallel checks (Code Quality, Build, Tests, Security, Type Safety). Max 3 retries then escalate. Event-driven failure routing.
Automatic test generation from code patterns and specs. Covers unit, integration, and E2E tests with Playwright integration.
Multi-agent orchestration between Trizai, OpenClaw, and Hermes Agent for complex cross-platform workflows.
Persistent conversation memory workflow — captures, consolidates, and injects learnings across sessions via the self-learning pipeline.
Multi-package monorepo patterns for shared code, dependency management, and coordinated releases.
Enterprise-scale architecture framework for large organizations with multiple teams, domains, and legacy systems.
Full project tracking system with milestones, epics, tasks, dependencies, and status reporting.
Docs-as-code workflow for API docs, ADRs, runbooks, and release notes with automated generation.
Team operating principles, communication standards, and collaboration patterns for consistent execution quality.
Task → role routing table mapping 30+ task categories to primary/backup/QA roles with formal handoff protocols.
Priority Stack
These are Trizai's primary technologies — deepest knowledge, first-class support, and priority focus for every project. Sorted by strength.
Utility-first CSS framework — responsive design, custom themes, design tokens
#1 PriorityProgressive frontend framework — SPA, Composition API, Pinia, Vite
#2 PriorityCross-platform UI toolkit — natively compiled iOS, Android, web, desktop from one Dart codebase
#3 PriorityCross-platform mobile framework — native iOS/Android, Capacitor, live reload
#4 PriorityPHP backend framework — Eloquent ORM, Artisan CLI, queues, Sail, Breeze
#5 PriorityVue meta-framework — SSR, SSG, file-based routing, auto-imports, modules
#6 PriorityRelational database — indexing, query optimization, replication, migrations
#7 PriorityTyped JavaScript — interfaces, generics, advanced types, strict mode
#8 PriorityGame engine & runtime — 2D/3D games, AR/VR, simulations, mobile games, C# scripting
#9 PriorityLocal Tools
Development tools on this machine that Trizai can access directly — launch, configure, build, and debug without leaving the conversation.
| Tool | Path | Status | Capabilities |
|---|---|---|---|
| VS Code | VS Code (local install) | v1.123.0 | Open files, folders, projects; edit & debug |
| XAMPP | XAMPP (local install) | Apache + MySQL + PHP 8.2 | Start/stop servers, PHP CLI, MariaDB 10.4 |
| DBeaver | DBeaver (local install) | v26.0.0 | Database GUI & CLI (dbeaverc.exe) |
| Android Studio | Android SDK (local) | ADB + Emulator + Gradle | Build, test, deploy, UI automation |
Integrations
Trizai is integrated with OpenClaw (multi-channel AI gateway) and Hermes Agent (self-evolving AI assistant). These tools extend Trizai's reach to mobile, messaging channels, voice, browser automation, self-evolution, and cross-platform deployment.
Multi-channel AI gateway with 50+ model providers, messaging integrations (Discord, Telegram, WhatsApp), browser automation, Docker/Nix deployment, and ClawHub skill registry.
18789 (loopback)openclaw (Node.js 22.19+)Self-improving AI assistant from Nous Research. Primary Discord gateway for Trizai. Features DSPy-driven self-evolution, skill curation, MCP server management, ACP protocol, web dashboard, and bidirectional memory sync with OpenCode.
9119hermes (Python venv)integrations/hermes/config.yamlintegrations/hermes/SOUL.md — Role Bundle SystemBidirectional memory synchronization between OpenCode and Discord. Every memory update in either platform syncs automatically via Hermes memory bridge. What Trizai learns in Discord, it knows in OpenCode, and vice versa.
Scheduler/scripts/hermes-memory-bridge.ps1Scheduler/scripts/watch-memory.ps1 (5-second interval)integrations/shared/hermes-memory.jsonAgent Communication Protocol bridge connecting OpenClaw Gateway to Hermes Agent. Enables multi-agent orchestration where Trizai routes tasks to the right tool.
C:\Triz\integrations\shared\acp-bridge.jsonC:\Triz\integrations\shared\workspaceMCP Servers
Trizai is wired with custom MCP servers that extend its reach into memory, projects, skills, GitHub, and Figma — all accessible through the OpenCode MCP framework.
Persistent memory operations — read, append, search, and manage ADRs across 7 memory files. Keeps Trizai context-aware across sessions.
Project indexing and discovery — list projects, read indexes, search across all TCode projects by component, route, or dependency.
Skill management — list, read, search, and navigate the full skill tree across 20 domains. Instant access to any skill definition.
Full GitHub API access — search code, repos, PRs; browse repo structures; get file contents. Direct integration via .mcp/github-mcp-server.
Figma design tool access — create frames, text, shapes; manage variables and styles; export screenshots and tokens. Integrates with .mcp/figma-plugin.
Dedicated skill file (Skills/general/custom-mcp/) with patterns for building, testing, and deploying custom MCP servers. Covers SDK setup, tool/resource patterns, and error handling.
Design System
A cohesive, dark-themed design system built around the Trizai monochrome brand — combining terminal aesthetics, bordered cards, and precision typography with JetBrains Mono font stack.
Pure monochrome — black (#000000) and white (#ffffff) with gray borders (#2a2a2a). Terminal-inspired aesthetic.
JetBrains Mono (300–800) for everything — clean, technical monospace for a terminal-inspired brand.
Glass morphism cards, gradient CTAs, tilt-shine hover effects, role cards, stat cards, progress bars, chips & tags.
Fade-in-up on scroll, hero 3D float, radial mesh shift, neon pulse glow, responsive breakpoints at 1024/768/420px.
Explore every color swatch, type scale, component, and animation in detail.
Stash
Trizai's persistent stash — saved items, links, notes, tips, and ideas stored in Memory/stash.md. Every entry is dated, tagged, and never deleted (only archived).
Everything you save stays forever. Entries are tagged (system, stack, tools, config, roles, skills, html, tips, admonition, todo, idea, reference, project).
Say Stash this: [item] with tag [tag] — Say Show my stash — Say Show stash tagged [tag] — Say Unstash #[number]
Starting Up
Download the Trizai folder to your machine. The entire system is portable — just copy the Triz/ directory anywhere.
Run npm install -g opencode-ai to install the runtime that powers Trizai. Requires Node.js 18+.
Run opencode providers set opencode-go <your-api-key> to authenticate. Trizai uses the model configured in opencode.json.
Run opencode C:\Triz from the command line. The harness in CLAUDE.md will load all modules and you are ready.