joao-aires

joao-aires/agent-skills-public

Skill Claude Code

Choose and evolve application architecture from product outcomes, existing system evidence and operating constraints. Use after product refinement for a new system or before planning a feature that changes service, data, client or integration boundaries; avoid architecture ceremony for small local edits.

joao-aires/agent-skills-public

Skill Claude Code

Assess and refine interface quality across hierarchy, typography, spacing, contrast, motion and component consistency. Use for a scoped UI critique or during substantial design, implementation and improvement; preserve accepted product direction and use rendered evidence for visual claims.

joao-aires/agent-skills-public

Skill Claude Code

Build and verify TypeScript Tauri desktop clients and iOS/Android clients connected to application backends. Use for native shell integration, accepted mobile frameworks, platform lifecycle, secure client storage, packaging and native testing; preserve established client frameworks.

joao-aires/agent-skills-public

Skill Claude Code

Plan task-based usability studies and analyze supplied observations for a specific interface or flow. Use for research questions, participant tasks, neutral discussion guides or evidence synthesis; distinguish proposed studies, heuristic review and actual participant findings.

joao-aires/agent-skills-public

Skill Claude Code

Review, write or harmonize interface labels, instructions, validation, errors, empty states and confirmations in their task context. Use directly for UI copy or during relevant design and implementation; preserve product behavior, established voice, localization and consequential meaning.

joao-aires/agent-skills-public

Skill Claude Code

Design, build and operate AI-enabled applications and agent integrations independently of a provider. Use for model/tool workflows, context and retrieval, skills, harnesses, MCP, agent interaction or generated UI, and their evaluation and delivery; apply technical guidance only to the accepted stack.

joao-aires/agent-skills-public

Skill Claude Code

Define and improve AI operating behavior across task observability, quality monitoring, latency/cost, provider limits, drift, rollout and recovery. Use for AI production readiness, operational diagnosis or learning from live behavior independently of a language, telemetry SDK or monitoring backend.

joao-aires/agent-skills-public

Skill Claude Code

Design and assess AI testing and evaluation across components, retrieval, agent trajectories, conversations, protocols and product journeys. Use for quality criteria, datasets, graders, regression gates and evaluation evidence independently of an implementation library or application language.

joao-aires/agent-skills-public

Skill Claude Code

Implement accepted Python model/provider adapters and standalone FastMCP interfaces. Use for Gemini integration, version-matched Python MCP APIs and interaction integration; preserve established providers and frameworks.

joao-aires/agent-skills-public

Skill Claude Code

Implement Python AI evaluation suites with DeepEval for model, retrieval, agent trajectory, conversation and end-to-end quality. Use when DeepEval is selected or being evaluated; preserve existing evaluation tools and do not introduce model calls, hosted reporting or synthetic generation without scope.

joao-aires/agent-skills-public

Skill Claude Code

Set up and verify Python AI observability with OpenTelemetry and compatible model/framework instrumentation. Use for AI task/tool/retrieval traces, usage and latency signals, export/privacy configuration and operational diagnosis; keep the chosen backend and quality evaluation separate.

joao-aires/agent-skills-public

Skill Claude Code

Define application-service behavior, authority, state and interface contracts independently of a framework. Use for APIs, durable jobs, stored data or integration changes before dependent implementation; preserve accepted behavior for local fixes.

joao-aires/agent-skills-public

Skill Claude Code

Configure selected GitHub Actions, Docker/Compose, local quality tools, telemetry backends and web test drivers. Use for their implementation and compatibility details after the project selects those tools; preserve existing alternatives.

joao-aires/agent-skills-public

Skill Claude Code

Derive an interface's visual direction from its users, tasks, content and environment. Use for a new product, substantial redesign, semantic tokens or a design-to-implementation handoff across web, desktop and mobile; preserve accepted identity during ordinary changes.

joao-aires/agent-skills-public

Skill Claude Code

Shape product direction from user problems, evidence and constraints. Use for discovery, prioritization, useful first slices, success measures and product sequencing before development or when new evidence changes priorities; reuse accepted decisions for defined changes.

joao-aires/agent-skills-public

Skill Claude CodeCodex

Builds ADK (Agent Development Kit) Python agents: LLM agents with tools, graph workflows of function and agent nodes, conditional routing, fan-out and join, schema-validated delegation between agents, human-in-the-loop pauses, and pytest coverage for all of it. Use when asked to create an agent or a workflow, add a…

joao-aires/agent-skills-public

Skill Claude CodeCodex

Explains how the ADK runtime fits together: the node and graph execution model, Context and Event flow, checkpoint and resume, tracing, and the rules governing the public API surface. Use when answering "how does X work" about ADK internals, tracing where an event or a piece of state comes from, deciding where a new…

joao-aires/agent-skills-public

Skill Claude CodeCodex

INVOKE THIS SKILL when setting up a new project or when asked about package versions, installation, or dependency management for LangChain, LangGraph, LangSmith, or Deep Agents. Covers required packages, minimum versions, environment requirements, versioning best practices, and common community tool packages for both…

joao-aires/agent-skills-public

Skill Claude CodeCodex

Create LangChain agents with createagent, define tools, and use middleware for human-in-the-loop and error handling.

joao-aires/agent-skills-public

Skill Claude CodeCodex

INVOKE THIS SKILL when writing ANY LangGraph code. Covers StateGraph, state schemas, nodes, edges, Command, Send, invoke, streaming, and error handling.

joao-aires/agent-skills-public

Skill Claude CodeCodex

INVOKE THIS SKILL when your LangGraph needs to persist state, remember conversations, travel through history, or configure subgraph checkpointer scoping. Covers checkpointers, threadid, time travel, Store, and subgraph persistence modes.

joao-aires/agent-skills-public

Skill Claude CodeCodex

Guide for creating high-quality MCP (Model Context Protocol) servers that enable LLMs to interact with external services through well-designed tools. Use when building MCP servers to integrate external APIs or services, whether in Python (FastMCP) or Node/TypeScript (MCP SDK).

joao-aires/agent-skills-public

Skill Codex needs its repo

Configures Python projects with modern tooling (uv, ruff, ty). Use when creating projects, writing standalone scripts, or migrating from pip/Poetry/mypy/black.

joao-aires/agent-skills-public

Skill Claude Code

Design or evolve stored data, constraints, transactions and migration behavior independently of a database engine. Use for persistence choices, schema changes, consistency boundaries and data lifecycle decisions.