techygarg/lattice

Architecture enforces architectural rules when generating or reviewing code, and checks proposed designs against loaded architecture standards. It validates layer responsibilities, dependency direction, and structural constraints for clean, hexagonal, modular monolith, or custom architectures.

Architecture Compass scans a repository, interviews the user about unclear parts, and drafts a shareable architectural assessment. Use it to understand the current structure, drift, and recommended direction for one codebase, module, or folder.

Architecture Refiner facilitates a structured conversation to define architecture principles for a repository. It supports clean architecture, hexagonal/ports and adapters, modular monolith, or a custom style when setting up or revising project standards.

Bug Fix investigates, reproduces, and safely repairs a bug with regression protection. It is used when fixing failures, regressions, or broken behavior in a codebase.

Clean Code applies clean code principles when generating or modifying implementation code. It helps with function focus, naming clarity, complexity management, error handling, and self-documenting style.

Clean Code Refiner guides a structured conversation to define clean code principles for a repository. It produces a clean-code.md document used as the override or overlay for coding standards and code quality rules.

Code Forge generates implementation code from an approved blueprint or verbal requirements. It is used when moving from design to code, implementing contracts, or responding to requests like “implement” or “build it.”

Collaborative Judgment defines when an AI should decide, when it should ask, and how to present genuine tradeoffs during code generation, design, and review. It is used when facts are missing, sources conflict, or multiple grounded approaches remain.

Context Anchoring manages per-feature living documents that capture decisions, constraints, and reasoning across AI sessions. Use it to create, load, or enrich context for a feature during design, implementation, bugfixes, or refactors.

Ddd Refiner facilitates a structured conversation to define DDD guardrails for domain design in a repository. It produces a ddd-principles.md document for use as an override or overlay when setting up domain rules or aggregate conventions.

Design Blueprint runs a structured design workflow from context setup through four design levels to an approved blueprint. Use it when planning architecture, starting a feature design, or resuming work from an existing context doc.

Design First guides structured thinking through five levels before coding: capabilities, components, interactions, contracts, and implementation. Use it for new features, refactors, module design, or any request to plan or architect before writing code.

Domain-Driven Design applies tactical DDD patterns to domain code and validates proposed models before approval. Use it when creating or changing aggregates, value objects, entities, or bounded context boundaries.

Knowledge Priming loads project-specific context like tech stack, architecture, directory layout, trusted sources, and conventions before other skills run. Use it when a knowledge base exists or when asking what the project uses and how it is organized.

Knowledge Priming Refiner facilitates a structured conversation to create a project-specific knowledge base for AI context. It is used when setting up or revising project context, tech stack details, architecture, trusted sources, and directory layout.

Language Idioms Refiner guides a structured interview to capture a project's language-specific idioms and patterns. Use it when setting up a new project, switching languages, or defining how pseudocode should map to Go, Rust, Python, and similar languages.

Lattice Init scans a project, detects existing Lattice configuration, and creates or reconciles .lattice/ settings. It also flags outdated requirements layouts and other optional upgrade gaps when you run lattice init or check for upgrades.

Lattice Sync audits Lattice documentation against the live skill inventory and fixes stale or missing references in core docs. Use it when skills change and you need README, docs, templates, and CLAUDE.md to stay in sync.

Learning Harvest manages operational learnings across sessions by loading prior patterns, capturing new ones, and keeping the document concise. Use it when a workflow session ends, when starting work that should benefit from past insights, or when harvesting repeated patterns.

Refactor Safely restructures existing code without changing externally visible behavior. It guides characterization-first refactors with context, architecture, code quality, and test guardrails.

Skill Enhance rewrites an existing Lattice skill to modern-capable-model grade with grammar fixes, deduplication, behavior preservation checks, QA, and independent verification. Use it when you want to upgrade or modernize a skill end to end.

Skill Forge creates new Lattice skill files for atoms, molecules, or refiners and follows the framework’s conventions. Use it when adding a new Lattice skill or when you need guidance on the required file structure and section order.

Skill Review audits a Lattice skill for realistic behavioral gaps by comparing three personas and merging only high-confidence findings. Use it after creating or changing a skill to find missing scenarios, silent failures, and inconsistencies.

Skill Tighten audits Lattice skill files for language compliance and rewrites phrasing that agents may underweight at runtime. It removes rationale prose, tightens soft language into imperatives, adds STOP gates to hard rules, and cuts redundant repetition.

Skill Validator checks a Lattice SKILL.md against tier conventions, cross-references, and structural rules. Use it when validating a new or edited skill file before it reaches the repo.