Monstrare — a gate-based workflow layer for AI coding agents. Every non-trivial change is spec'd, planned, verified, and reviewed before it ships — nothing is called done without evidence.
Skill Claude CodeCodex
A set of visual design rules for building and reviewing user interfaces, covering typography, spacing, colours, and comparison with reference projects.
Skill Claude CodeCodex
A planning tool that turns an approved software specification into architecture notes and small task cards for an AI coding assistant. Each task includes limits and checks for confirming the result.
Skill Claude CodeCodex
A workflow for creating a new project’s work list in layers: large deliverables, user stories, and individual tasks. Each layer is reviewed by a person before being added to a project-management board.
Skill Claude CodeCodex
A skill for searching a code repository with a small number of targeted searches and assembling task-specific context. It can also update a lasting map of the project when needed.
Skill Claude CodeCodex
A review procedure for checking plans or code changes for correctness, security, privacy, maintainability, and architectural consistency.
Skill Claude CodeCodex
A requirements-review workflow that turns vague feature requests into a written specification. The specification lists requirements, exclusions, edge cases, and acceptance criteria for human review.
Skill Claude CodeCodex
A skill for planning, running, and collecting evidence from software checks. It covers tests, builds, lint checks, security checks, and UI screenshots.
Skill Claude CodeCodex
A review gate that creates several UI mockup variations and a document describing screen states before implementation. It then pauses for a person to choose a direction.
At most 3 mods per repository are shown here, and a mod shipped inside a plugin is left to that plugin's page — the rest are on their repository pages: