Bilingual (English · Español) wiki about the engineering layers needed to work with AI and build agentic systems.
Una wiki viva con un pipeline autónomo diario: un collector RSS/Atom (Python) alimenta un gate de decisión (Jev vía OpenRouter) y sólo los candidatos relevantes llegan a research generativo y a un PR revisable. La automatización propone, los humanos hacen merge, y GitHub Pages publica desde main.
sources/ — feed allowlist (feeds.yaml + JSON Schema), topic taxonomy, trust model.scripts/ — collector → normalize → deduplicate → decide (Jev) → retrieve → research → patch → l10n → verify → report → parity → model_verify.data/ — committed operational state: Source Memory, checkpoints, per-run reports.docs/{en,es}/ — canonical (EN) and mirrored (ES) site, built with two standalone MkDocs Material configs.python -m venv .venv
.venv/bin/pip install -r requirements.txt # Windows: .venv\Scripts\pip
python -m pytest
python -m ruff check scripts tests
python scripts/parity.py
python -m mkdocs build -f mkdocs.en.yml --strict
python -m mkdocs build -f mkdocs.es.yml --strict
See Setup checklist for the GitHub configuration (Pages → GitHub Actions, OPENROUTER_API_KEY secret, model variables).