Frederico Guilherme Klüser de Oliveira
Community@frederico-kluser · São Paulo, SP, Brazil
Engenheiro de Software
Agent Skills by Frederico Guilherme Klüser de Oliveira
Showing 14 vetted skills indexed across 2 GitHub repositories.
frederico-kluserextending-the-crawler
Guides adding sources, modules, CLI commands, and fetch strategies to the newsletter crawler.
frederico-kluserfollowing-code-style
Enforces ESM, logging, naming, and SQL conventions when writing JavaScript in this repository.
frederico-klusersearching-the-corpus
Search a crawled newsletter corpus with LLM relevance judging and generate PT-BR summaries.
frederico-kluserbuilding-the-ink-tui
Builds and extends Ink/React terminal menus with htm, theme tokens, and live progress panels.
frederico-kluserfetching-and-extracting
Guides HTML fetching, rendering, and article extraction decisions in the newsletter crawler codebase.
frederico-klusermeta-skill-evolution
Decides whether to update, propose, or discard skill knowledge after each task.
frederico-kluserpersisting-and-orchestrating
Documents SQLite schema, frontier state machine, and crawl orchestration for a newsletter crawler.
frederico-kluserproject-router
Routes implementation tasks to the correct skills and manages the TASK_PLAN.md lifecycle.
frederico-kluserbuilding-the-webapp
Build and extend the static Vite React webapp frontend with bilingual i18n and theme tokens.
frederico-klusermeta-skill-consolidate
Deduplicates, revalidates, and prunes skill content under token budgets with regression gating.
frederico-klusercalling-the-llm-layer
Configure OpenRouter and DeepSeek LLM calls with structured output, retries, and cost control.
frederico-kluserrunning-and-verifying-crawls
Runs, smoke-tests, and verifies the newsletter crawler end to end via npm commands.
frederico-klusersurf-search-skill
Route web search and extraction across Tavily, Parallel AI, and Brave with key rotation.
frederico-klusersurf-plan-skill
Generate research-grounded coding execution plans with cited web sources.
Frequently Asked Questions About Frederico Guilherme Klüser de Oliveira
FAQPage SchemaWhat tasks can I accomplish with frederico-kluser's skills?▼
You can extend and operate a newsletter crawler: add sources and fetch/extract strategies, run bounded crawls with resume and dedup, search the saved corpus in two relevance modes, generate PT-BR summaries, build Ink terminal menus, and maintain the static Vite/React webapp frontend.
Who are these skills designed for?▼
Software engineers maintaining this specific Node.js crawler repository. They suit contributors editing src/ modules, adding CLI commands or UI screens, changing the SQLite schema or frontier queue, and reviewers enforcing the repo's ESM, logging, and naming conventions.
How do I run and verify the crawler in practice?▼
Use npm run status to boot the schema, npm test for the suite, and npm run crawl with --max-pages/--max-articles/--since bounds. Verify fixes via cache-hit re-runs, DB status counts, markdown export, the web UI check, and the OpenRouter key probe.
What are the prerequisites and dependencies?▼
The stack requires Node.js with ESM, SQLite via the db.js prepared statements, Playwright for rendered fetching, Ink/React for the terminal UI, and Vite/React for the webapp. The classification layer needs an OpenRouter or DeepSeek key set via .env or LLM_PROVIDER.
How is content extraction and politeness handled?▼
fetchSmart decides between static fetch and Playwright rendering with profiles, smart infinite-scroll, and date-floor stops. Extraction uses Readability in a child-process pool, DOM pruning, robots politeness, a circuit breaker, anti-bot interstitial rejection, and CSS selector validation thresholds.