lglucas/ai-dev-operating-system
Hook Claude Code
Runs before the agent uses a tool for Bash, Write, Edit, MultiEdit and NotebookEdit tool calls, executing block-secret-commit.js and protect-env-files.js via node (2 commands). From lglucas/ai-dev-operating-system.
lglucas/ai-dev-operating-system
Settings file Claude Code
Agent settings declaring 1 hook event (PreToolUse).
lglucas/ai-dev-operating-system
Skill Claude CodeCodex
Check every Business Plan change for downstream impact before accepting it — price moves projections, ARPU, LTV and positioning; persona moves GTM, UX and MVP scope; stack moves costs, roadmap and risks. Runs at WIZARD stage 2.8 when turning BP v0.0.1 into v0.0.2, and any time the BP is edited after approval. Use when…
lglucas/ai-dev-operating-system
Skill Claude CodeCodex
Mantém o CODEMAP.md — o índice de todo arquivo de código do projeto com uma linha sobre o núcleo de cada um, para achar arquivo sem ler o repositório inteiro. Regenere sempre que a lista, a descrição ou a contagem de linhas mudar — na prática, junto com qualquer edição de código —, ao fechar sprint e antes de release.…
lglucas/ai-dev-operating-system
Skill Claude CodeCodex
Flag potentially expensive technical choices BEFORE they are committed — AI model selection, hosting tier upgrades, third-party services with metered pricing, and DB patterns that explode with scale. Preventive counterpart to usage-monitor, which tracks real spend after launch. Triggers when the user says "posso usar…
lglucas/ai-dev-operating-system
Skill Claude CodeCodex
At the start of a session (or on user request "o que a gente fez", "onde paramos"), produce a 4-bullet briefing — what's done, what's in progress, what's next, what's blocked. Replaces the dev habit of standup for solo vibe coders.
lglucas/ai-dev-operating-system
Skill Claude CodeCodex
Record WHY a decision was made — including the alternatives rejected — as a session-log entry or an ADR. Use after any decision touching architecture, product scope, pricing, compliance, security, or release strategy, when multi-ai-review reconciles a disagreement, and when the user says "anota isso", "registra essa…
lglucas/ai-dev-operating-system
Skill Claude CodeCodex
Walk a vibe-coder through deploying a Next.js + Supabase project to Vercel + Supabase production for the first time. Activated when the user says "vamos publicar", "como coloca no ar", "deploy", or after Sprint 1 ships and the app needs a public URL.
lglucas/ai-dev-operating-system
Skill Claude CodeCodex
Create a feature-based folder structure (components, actions, api, db, hooks, schemas, tests) plus an initial implementation plan, following .claude/rules/feature-based-architecture.md. Use when starting a new product feature during a coding sprint, and when the user says "vamos criar a tela de X", "quero adicionar…
lglucas/ai-dev-operating-system
Skill Claude CodeCodex
Get the first 100 real users for a vibe-coder SaaS without spending on ads. Hand-curated outreach, founder communities, niche-specific channels, retention check-ins. Activated after launch when the user asks "como pego os primeiros usuários", "ninguém tá usando", "como divulgo isso?", or after launch-agent Track A.…
lglucas/ai-dev-operating-system
Skill Claude CodeCodex
Plan post-first-100-users growth without burning out the founder, the budget, or the user base. Pairs with first-100-users (which gets you to 100); this skill takes you from 100 to 1000 sustainably. Activated when the user says "já tenho 100 usuários, e agora?", "como cresço sem queimar dinheiro?", "o WhatsApp 1-a-1…
lglucas/ai-dev-operating-system
Skill Claude CodeCodex
Get a second (and third) AI opinion on hard-to-reverse decisions. Runs the same prompt through different agent roles or model families in parallel, then reconciles disagreements. Inspired by gstack's /codex cross-review pattern. Activated for architecture decisions, security-sensitive code, pricing/packaging changes…
lglucas/ai-dev-operating-system
Skill Claude CodeCodex
Verify the AI Dev Operating System is in a coherent state inside a project. Detects missing canonical files, broken cross-references, skills/agents/commands without frontmatter, registry drift, unindexed session logs, unwired hooks and gitignore gaps. Run after major edits to the OS, after renaming or moving canonical…
lglucas/ai-dev-operating-system
Skill Claude CodeCodex
Derive a 10-section pitch from BP v0.0.2 into docs/business/PITCH.md, then ask the founder whether the BP and Pitch should live online inside the product (/pitch, /investors) — running a redaction gate over what must never go public. Runs at WIZARD stage 2.9. Use when the user says "preciso de um pitch", "vou…
lglucas/ai-dev-operating-system
Skill Claude CodeCodex
Translate technical decisions, architecture, errors, or AI-generated jargon into plain Portuguese a vibe coder (non-dev) can act on. Use whenever the user signals confusion ("não entendi", "o que isso quer dizer", "explica de novo") or after producing technical artifacts.
lglucas/ai-dev-operating-system
Skill Claude CodeCodex
Review any feature that collects, stores, processes, exports, or displays personal data — answering the nine questions in .claude/rules/privacy-audit.md (what data, why, where, who can access, how controlled, retention, logs, deletion/export, policy impact). Mandatory before merging personal-data changes. Triggers on…
lglucas/ai-dev-operating-system
Skill Claude CodeCodex
Turn a manually-validated workflow into a documented, repeatable, and partially-automatable process. Use AFTER a task has been done at least 3 times by hand and is showing signs of repetition. Inspired by Sahil Lavingia's principle "do it manually until it works, then processize." Activated when the user says "tô…
lglucas/ai-dev-operating-system
Skill Claude CodeCodex
Write docs/product/PRODUCT-BRIEF.md by reverse-engineering the approved prototype — every screen, the flow as actually clicked, and the gaps the prototype implied but never showed (empty states, errors, permissions, offline). Runs at WIZARD stage 4.1, after BP v0.0.2 and after a prototype direction is approved in…
lglucas/ai-dev-operating-system
Skill Claude CodeCodex
Drive the full 5-phase Project Genesis Wizard, from raw idea to first coding sprint. Use when starting a new project from the AI Dev Operating System, when the user runs /project-start, or when they say "vamos começar um projeto novo", "tenho uma ideia de SaaS", "quero criar um app", "por onde eu começo?". Enforces…
lglucas/ai-dev-operating-system
Skill Claude CodeCodex
Build three genuinely distinct HTML/CSS/JS prototype directions — brand, color, typography, UI, UX — from BP v0.0.2 and the research, then extract the approved one into docs/product/DESIGN-DIRECTION.md. Runs at WIZARD Phase 3, BEFORE the Product Brief and Technical Plan. Use when the user says "quero ver como vai…
lglucas/ai-dev-operating-system
Skill Claude CodeCodex
Recommend external repo packs from docs/registry/ that fit a project's stack, domain, and constraints. Runs twice in the WIZARD — a design-scoped pass at stage 3.1 (before the Prototype Lab) and a full stack pass at stage 4.3 (after the Technical Plan). Also use when the user asks "que repos eu deveria usar?", or…
lglucas/ai-dev-operating-system
Skill Claude CodeCodex
Pre-release gate covering tests, build, docs, changelog, privacy, security, and sprint closure. Delegates to verify-build-works, secrets-scan, and privacy-audit rather than re-checking by hand. Run before tagging a release, before closing a sprint, before a first deploy, and when the user says "tá pronto pra lançar?"…
lglucas/ai-dev-operating-system
Skill Claude CodeCodex
Run the 3-wave research and validation process — Wave 1 competitor and market research, Wave 2 red team (devil's advocate, business/finance, technical/security), Wave 3 coordinator consolidation and BP writing. Runs at WIZARD stages 2.4–2.6, only after the user approves the plan. Use when the user asks "quem são meus…
lglucas/ai-dev-operating-system
Skill Claude CodeCodex
Undo the last AI change without losing user work. Use when the latest AI-applied change broke something, the user says "desfaz", "volta", "undo", "estraga tudo", or after bug-triage-agent selects a revert as the safest fix. Designed for vibe coders who don't yet trust git.