fgiova
- 8 skills
- 0 followers
- 5 hours ago last updated
- ▌ Fastify Expert · fgiovaALWAYS consult this skill when the user mentions Fastify, builds Node.js APIs, creates routes or plugins, uses TypeBox/typebox for schema validation, configures env-schema, sets up OpenAPI/Swagger, or works on any Fastify-based backend — even for seemingly simple questions. This skill contains critical reference architecture, TypeScript patterns, and production-ready code templates that Claude cannot generate correctly without consulting. Covers: plugin architecture with fp(), decorators, schema validation, project structure, configuration, graceful shutdown, Dockerfile, autoload, Redis integration, Sentry setup. Does NOT cover test code (see fastify-testing) or container infrastructure (see tap-testcontainers). Trigger on ANY mention of: fastify, fastify plugin, route, schema, typebox, decorator, fp(), pino, autoload, close-with-grace, env-schema, @fastify/*, swagger, openapi, Node.js API, REST API, microservice backend, Fastify instance, server.ts, config.ts.
- ▌ Fastify Testing · fgiovaALWAYS consult this skill when the user writes, modifies, or asks about tests for Fastify applications using node-tap. This skill contains mandatory patterns for test structure, TypeScript types (TestWithContext, BuiltPlugins), assertion methods, naming conventions, MockAgent setup, Prisma test lifecycle, and database data management that Claude cannot produce correctly without consulting. Without this skill, Claude will generate incorrect test types, wrong lifecycle hooks, and missing mock cleanup patterns. Covers: tap test structure, t.test/t.equal/t.same/t.match/t.has/t.rejects, app.inject, beforeEach/afterEach patterns, undici MockAgent with dispatcher save/restore, test naming, TestWithContext typing, Prisma connection lifecycle, test data management, t.plan usage. Does NOT cover container infrastructure (see tap-testcontainers) or app architecture (see fastify-expert). Trigger on ANY mention of: test fastify, tap test, write tests, add tests, app.inject, t.test, t.equal, t.same, t.match, MockAgent, undi
- ▌ Prompt Engineer · fgiovaStructured prompt engineering workshop. Guide user through 6-phase interactive process to build pro-grade prompts: (1) Task definition with success criteria, (2) Context file creation with standards/constraints/audience, (3) Reference analysis + reverse-engineer into Always/Never rules, (4) Success brief with output specs, (5) Guardrail wiring, (6) Collaborative refinement instructions. Produce complete, copy-ready prompt in English with optional clipboard export via pbcopy. Without skill, prompt creation skip critical phases like context file gen, reference pattern extraction, structured success briefs — produce one-shot prompts instead of proven 6-phase methodology. MUST consult for: build prompts, create prompt templates, prompt engineering, design AI prompts, craft reusable prompts, improve existing prompts, reverse-engineer reference docs into prompts. Italian: crea un prompt, scrivi un prompt, prompt strutturato, voglio un prompt, aiutami con un prompt, migliora questo prompt.
- ▌ Localstack Expert · fgiovaActivate when working with LocalStack, setting up local AWS development environments, configuring docker-compose for AWS services, or using cdklocal/awslocal tooling. Trigger keywords: localstack, local aws, awslocal, cdklocal, localstack docker, localstack compose, localstack init, localstack lambda, localstack s3, localstack sqs, localstack dynamodb, localstack sns, localstack kinesis, localstack stepfunctions, localstack eventbridge, localstack secrets manager, localstack ses, localstack api gateway, local cloud, AWS_ENDPOINT_URL, LOCALSTACK_AUTH_TOKEN, docker-compose aws.
- ▌ Code Implementation · fgiovaEnd-to-end code delivery for complex implementation work: plan, build, review, document. Use when user want to actually IMPLEMENT feature, service, or refactoring - not just plan - and want result reviewed and documented. Activates for: "implement", "build", "develop", "realizza", "implementa", "sviluppa", "code this feature", "build and document", or any request to deliver working code with review and documentation. Runs full pipeline: decompose into plan, execute tasks, run adversarial code review over what built, then write English docs (README.md in project root and topic-split docs under docs/). Planning ONLY (no build): use prompt-decomposition instead.
- ▌ Tap Testcontainers · fgiovaALWAYS consult this skill when the user works with Testcontainers, Docker containers for tests, tap before/after hooks for container lifecycle, or test environment setup with test-env.json and localtest.ts. This skill contains the exact architecture for container-based test infrastructure (Ryuk reaper, before.js executors, container runners, teardown.js, localtest.ts environment loader) that Claude cannot reconstruct correctly without consulting. Without this skill, Claude will produce incorrect container lifecycle management, miss reaper label configuration, and break inter-process environment variable passing. Covers: Testcontainers setup, GenericContainer, Ryuk reaper, before.js/teardown.js executors, container runners (Redis, MySQL, PostgreSQL, LocalStack), test-env.json, localtest.ts, TEST_LOCAL and SKIP_TEST_* environment variables, tap before/after hooks for containers. Does NOT cover test code patterns (see fastify-testing) or app architecture (see fastify-expert). Trigger on ANY mention of: testconta
- ▌ Prompt Decomposition · fgiovaStructured planning and task decomposition for complex, multi-step work. Use this skill whenever the user needs to break a large task into smaller parts, create an implementation plan, organize a migration, or structure a refactoring. Activates for: planning complex features, migrating architectures, refactoring codebases, organizing multi-service work, or any request that involves multiple phases of coordinated implementation. Also handles resuming or suspending previously saved plans. Without this skill, complex requests are executed monolithically instead of being decomposed into atomic, parallelizable sub-tasks with dependency tracking.
- ▌ Platformatic Runtime · fgiovaALWAYS consult this skill when the user works with Platformatic Runtime, watt.json, the create() function pattern, inter-service communication via plt.local, or any project using @platformatic/* packages. This skill contains critical patterns for multi-service composition, runtime configuration, service entry points, and testing strategy that Claude cannot produce correctly without consulting. Without this skill, Claude will incorrectly call app.ready() inside create(), hardcode service URLs instead of using plt.local, and misconfigure watt.json environment variable mapping. Covers: watt.json configuration (dev and production), create() function pattern, inter-service communication via plt.local, pnpm workspace monorepo structure, service configuration with env-schema, unit vs integration testing strategy, common pitfalls. Trigger on ANY mention of: platformatic, watt, watt.json, plt, plt.local, @platformatic/*, create() function, platformatic runtime, platformatic service, platformatic node, entrypoint, watt