This document outlines our application's architectural principles and structure. It focuses on how code is organized at a high level, how components interact, and how to maintain clear boundaries between different parts of the application.
ClawQL production code that can be Effect-based must be Effect-based. There is no “pure sync is fine” exception and no standing exception for Promise domain APIs.
Use whenever code touches newsletter, marketing-email, drip-campaign, re-engagement, win-back, lead-magnet, or any outbound email with promotional content — even if the user's prompt doesn't mention DSGVO, GDPR, UWG, privacy, or "personal data". Triggers on Mailchimp / @mailchimp/mailchimpmarketing, Brevo /…
A set of operating rules for cc-master, a coordinator that manages long-running coding goals and delegates individual tasks. It is intended to direct the coordinator rather than perform each task itself.
Optimize and rewrite prompts using proven prompt engineering techniques. Invoke when the user wants to improve, optimize, or rewrite a prompt, or pastes a prompt draft asking to make it better.
Behavioral guidelines to reduce common LLM coding mistakes. Use when writing, reviewing, or refactoring code to avoid overcomplication, make surgical changes, surface assumptions, and define verifiable success criteria.
Musk 5-step engineering algorithm. Primarily for reviewing and cleaning up existing systems. Apply in strict order: question requirements, delete parts, simplify, accelerate, automate. Bias toward deletion. Never advance to a later step until earlier steps are visibly complete.
When tickets have been assigned or revising the implementation plan.
★not rated 12 1y agoA11 tokens
originalMIT
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: