alebgl77
UserThe self-distilling model router for AI coding CLIs — solve each problem class at the expensive tier once, execute it cheap forever. Just markdown, no proxy.
Categories
Indexed Skills (5)
model-routing
Complexity-based routing of tasks across Fable 5 (T3), Opus 4.8 (T2), Sonnet 4.6 (T1) and Haiku recon (T0). Use when planning any multi-step work, when deciding which subagent should execute a task, when writing a dispatch brief, or when an executor fails and escalation is needed.
tier-decay
Post-completion distillation protocol that converts Opus/Fable solutions into cheaper-tier playbook entries and recalibrates routing from real outcomes. Use after completing any [T2] or [T3] task, when logging a task outcome to the routing ledger, when a playbook probe passes or fails, or when a class keeps escalating.
grafana-llmops-forge
Pilote l'observabilité IA/LLM de bout en bout sur Grafana (OSS, Cloud, Enterprise) avec un seul prérequis — une instance Grafana accessible. Auto-découverte (version, APIs, datasources), détection des signaux LLM présents (OTel GenAI gen_ai.*, LiteLLM, vLLM, TGI, GPU/DCGM), génération et déploiement automatiques de dashboards — FinOps & coûts multi-providers (US/EU/Asie), gateway, tracing agents & RAG, adoption interne, gouvernance EU AI Act, inference self-hosted — alerting SLO, registre de modèles avec pricing, puis contrôle visuel du rendu réel (captures renderer/navigateur inspectées par vision). Utilise ce skill dès que l'utilisateur mentionne Grafana, un dashboard, du monitoring/observabilité IA ou LLM, des coûts de tokens, du FinOps IA, du LLMOps, des agents ou du RAG, l'adoption de modèles, la conformité AI Act, ou Prometheus/Loki/Tempo appliqués à l'IA — même sans le mot « dashboard ». Vaut aussi pour auditer un Grafana existant ou une stack IA qui n'émet encore rien.
execution-standards
Shared implementation discipline for executor agents — conventions, testing, diff hygiene, reporting. Preloaded into executor and heavy-executor via their skills frontmatter.
repo-playbook
Repo-specific execution patterns distilled from higher-tier (Opus/Fable) solutions via the tier-decay protocol. Preloaded into executor and heavy-executor at startup. Written ONLY by the main thread — executors read, apply, and report; they never edit this file.
Bio shown is the top-scored skill's repo description as a fallback — real GitHub bios land in a future update.