Software Engineer · Native AI Engineering · Skills Architecture · Product Design Systems

Building reusable AI-native engineering capabilities—turning methods into composable agent behavior with explicit gates and verified outcomes.

I architect frameworks and executable skills that enable consistent, evidence-backed engineering workflows. Focus: contracts, ports, behavioral validation, and seamless agent-team composition.

Featured Work · Operating Principles · Quick Start · Connect

90+ reusable skills, 10 workflows, 6 meta-skills with explicit phase gates and behavioral evaluation.

Creator & maintainer · Open source · Actively maintained

Transforms engineering, product, design, review, and delivery practices into bounded, composable capabilities. Each skill declares scope, procedure, evidence requirements, and quality gates. Works with Claude, Hermes, Cursor, Copilot, or any Agent Skills-compatible runtime.

• workflow-router — Let the agent choose the right lifecycle
• project-instruction-generator — Scoped, standardized agent instructions
• implementation-context-discovery — Repository frameworks and reuse patterns
• code-review-workflow, deployment-workflow, redesign-workflow — Ordered lifecycles with gates
• design-audit, decision-provenance, skill-doctor — Verification & review

Install: npx skills add puterakahfi/ai-native-skills -g -y

Runtime-agnostic contracts and domain abstractions for AI-native engineering.

Framework designer · Open source · Public specification

Defines canonical concepts, lifecycle rules, ports, adapter boundaries, and quality standards. Language-agnostic—implemented across Go, Python, Node.js, and beyond.

Orchestration, adapters, context discovery, and product runtime behavior.

Framework builder · Open source

Bridges reusable skills with product systems. Handles agent discovery, context enrichment, skill installation, runtime binding, and control-plane behavior.

AI Creative Control System for structured, reviewable creative workflows.

Product builder & system designer · Product · Private implementation

Connects product direction, workflow architecture, design systems, and AI-assisted creative production. Demonstrates practical Native AI patterns in real product delivery.

AI agents should improve the engineering system—not only generate more artifacts.

Purpose before decoration.
Proof before badges.
One narrative across all surfaces.
Explicit boundaries before hidden magic.
Evidence before feature count.

• Explicit scope & boundary — Every capability declares what it owns and what it composes
• Evidence-backed decisions — Decisions are traceable, reviewable, and have supporting proof
• Composable architecture — Skills route to skills; workflows compose phases with gates
• Contract-driven contracts — Ports, adapters, and runtimes stay decoupled; behavior is verifiable
• Behavioral validation — Regression tests protect reusable learning; quality gates prevent drift

Current focus · July 2026

Refining Native AI Engineering into a practical, battle-tested public framework while exploring reusable agent design patterns across UX, visual systems, identity, and constrained documents.

• Software architecture — Domain-driven design, ports-and-adapters, API contracts, event-driven systems
• AI-native orchestration — Agent roles, skill composition, context discovery, behavioral gates, runtime binding
• Product engineering — Go, Python, TypeScript/Next.js, Docker, modular systems, developer experience
• Design systems — Hierarchy, adaptive components, cross-medium design, evidence-backed review

npx skills add puterakahfi/ai-native-skills@workflow-router -g -y
npx skills add puterakahfi/ai-native-skills@project-instruction-generator -g -y
npx skills add puterakahfi/ai-native-skills@implementation-context-discovery -g -y
npx skills add puterakahfi/ai-native-skills@code-review-workflow -g -y

npx skills add puterakahfi/ai-native-skills -g -y

npx --yes skills add puterakahfi/ai-native-skills \
  --skill hermes-profile-bootstrap \
  --agent hermes-agent --global --yes

hermes chat -s hermes-profile-bootstrap -q \
  "Bootstrap the ai-native-engineering profile using the engineering preset."

• Skill Registry — Complete taxonomy
• Capability Discovery — Browse by domain, lifecycle, concern
• Skill Packs — Pre-composed workflows
• Contributing — Add a skill, reference, or eval case

ai-native-core
  ↓ canonical domain, contracts, terminology, quality standards
  ↓ implemented as executable behavior
ai-native-skills
  ↓ 90+ reusable skills, 10 workflows, 6 meta-skills
  ↓ installed directly or orchestrated through adapters
native-ai-fw & product repositories
  ↓ runtime execution, control-plane behavior, product validation

All surfaces tell one compatible story: identity → proof → selected work → current focus → action.

Open to conversations around:

• Software architecture & design systems
• Native AI Engineering & agent-driven development
• Product systems & developer experience
• Behavioral evaluation & verification patterns

Find me:

• GitHub — Code, issues, discussions
• pkahfi.com — Writing and archive
• GitHub Discussions — Ideas and collaboration
• VisualMate — Product work