Low-level agentic design patterns. Turning old engineering wisdom into reflexes your agent reaches for on its own—on any agent.

Make the most out of your subscriptions. Delegate work to other harnesses and models, while keeping your main session the orchestrator. Nothing new to learn. Keep working like you do, but better.

Master prompts to make Claude (and other AI models) review code for security, performance, accessibility, usability and quality — language-agnostic, with add-ons for 18+ stacks. Bilingual EN/ES.

Autonomous high-reasoning operating mode for Claude Code — Fable 5's working behaviors as a drop-in skill

Claude Code skill that clones Fable 5's reasoning habits — the playbook Fable wrote about itself. A/B validated: 20/20 convergence, blind-judge parity. One-line install.

Personal agent operating system

Claude Code skill + verifier for running Fable 5 background subagents: avoid silent Opus fallback, verify the real model, and structure slow hard work.

Unlock the hidden Claude Fable 5.1 model on macOS

Field-tested operating layer that transfers Claude Fable 5's working discipline to Claude Opus 4.8 via project instructions — with the eval harness that validated it (4 rounds, 46 artifact-verified runs).

Make Opus work like Claude Fable 5 - a Claude Code plugin that injects Fable's working doctrine at session start, with per-task playbooks.

Claude Code plugin installing the Fable Way working discipline: a Stop-hook calibration gate (done-claims must carry their evidence ledger), runner skills for scope/debug/review/verify/ship, and self-building per-project memory (.fable/). Built for Opus/Sonnet