بنقرة واحدة
context-builder-kit
يحتوي context-builder-kit على 6 من skills المجمعة من j4th، مع تغطية مهنية على مستوى المستودع وصفحات skill داخل الموقع.
Skills في هذا المستودع
Make stack and methodology decisions, produce the foundation doc set, and define the initiative for an AI-assisted development project. Use this skill whenever the user has a problem brief and scaffold output and is ready to make strategic technical decisions. Triggers include "let's blueprint this", "time for stack decisions", "set up CLAUDE.md and standards", "what stack should I use", "I need a CLAUDE.md", "how should I architect this", "I have my repo set up, what's next", "define the initiative", or any reference to the cascade at the blueprint level. Also use when a user skipped earlier phases — the skill will note missing inputs and offer to run them. **Use this skill even when the user doesn't explicitly say "blueprint" — the trigger is the moment of "I have a workspace and need to decide how to actually build this", not the vocabulary.** Phase 3 of the cascade. Produces six prose docs plus tooling configs and a handoff issue, and either creates a Linear initiative or commits blueprint.md to the repo.
Turn a raw idea into a bounded problem brief before any workspace, repo, or planning artifact exists. Use this skill whenever the user has an idea they want to build, even vaguely phrased — "I want to make X", "I've been thinking about building Y", "I have this idea but don't know where to start", "where do I begin with X", "is this even worth doing", "help me think through Z". Also use when the user asks to "shape", "scope", or "frame" a project, mentions the planning cascade, or describes a problem they want to solve without yet naming a solution. **Use this skill even when the user doesn't explicitly ask for "consultation" or "a problem brief" — most users at the consultation moment don't know the cascade exists, so the trigger is the shape of the conversation, not the vocabulary.** Phase 1 of the six-phase AI-assisted development cascade. HITL-heavy and produces no code, no repo, no tickets — only a structured problem brief.
Decompose a Linear project into sequenced milestones with rough issues. Use this skill when the user wants to break a project into milestones, plan implementation for a project, sequence build order, or move from planning into execution. Trigger when the user says 'plan this project', 'frame this project', 'what are the milestones', 'break this project down', 'how should I sequence this work', 'I have my plan, what's next', references a specific project by name, or references the cascade at the framing level. Also trigger when the user picks a workstream from blueprint and wants to go deeper. **Use this skill even when the user doesn't explicitly say "framing" — the trigger is the moment of "I have a workstream from blueprint and need real milestones I can build against", not the vocabulary.** Phase 4 of the cascade. One project at a time, just-in-time, never all at once. Produces a numbered cascade-event file (frame-NN.md) plus updates the chronological framing index.
Decompose one framing milestone into Claude-Code-ready sub-sub-issues. Use this skill when the user wants to break a milestone into implementable issues, move from planning into execution, produce issue specs with acceptance criteria and Claude Code plan-mode prompts, or pick up the next milestone for implementation. Trigger when the user says 'rough in this milestone', 'break this milestone down', 'plan issues for M1', 'I have my framing, what's next', 'let's start building', references a specific framing milestone by F-number, or references the cascade at the rough-in level. **Use this skill even when the user doesn't explicitly say "rough-in" — the trigger is the moment of "I have a framing and need specs I can hand to Claude Code", not the vocabulary.** Phase 5 of the cascade. One milestone at a time, just-in-time. Produces ready-to-implement sub-sub-issues under a framing sub-issue, each with acceptance criteria, technical detail, and a Claude Code plan-mode prompt.
Provision the workspace infrastructure (repo, planning surface, knowledge surface) that the rest of the AI-assisted development cascade will use. Use this skill whenever the user has a problem brief and needs to set up where the work will live, or says things like "set up the workspace", "bootstrap the repo", "I'm starting a new repo", "where should I put this project", "I need to set up GitHub for this", "I have my plan, now what", or references the cascade at the scaffold level. Also use when a user wants to start a new project end-to-end and skipped consultation. **Use this skill even when the user doesn't explicitly ask for "scaffolding" or "bootstrap" — the trigger is the moment of "I have a plan and need a place to put it", not the vocabulary.** Phase 2 of the cascade. Supports three backend profiles — GitHub-only (default), opinionated Linear+Notion, and markdown-only (no planning surface, for users who don't want a board). Produces a scaffold output doc and bootstrap checklist.
Create a new immutable ADR in docs/adr/, allocating the next sequence number, rendering from template.md, and updating the three indexes (docs/adr/README.md, docs/ARCHITECTURE.md § Decisions log + Configurability summary, docs/cbk/blueprint.md § Stack decisions). Use when the user wants to record a new architecture decision or supersede an existing one.