Runners‑and‑Cartridges

Runners and Cartridges

Runners and cartridges are optional modules that extend the Genesis Code Protocol with domain expertise, evaluation metrics and tools. They allow the base protocol to adapt to a wide range of fields without modifying its core logic.

Runners

Runners are deterministic modules that encapsulate domain expertise for specific phases. Examples include Code, Hardware, Theory, Operational Technology (OT), Mobility, Life Sciences, Agriculture/Measurement, Reporting & Verification (AgMRV), FinTech, Space, Deep‑Sea and Exotics. Each runner defines its own inputs, steps, outputs, gate criteria, prompts, telemetry and observability controls. Attach a runner when your Spark or project requires specialised knowledge or toolchains.

Cartridges

Cartridges are smaller, attachable modules that provide metrics, policies, tests and lexicons for a domain. Each cartridge bundles embeddings, indexes, vocabularies, evaluation hooks and optional tool contracts. Cartridges support retrieval‑augmented generation (RAG), metamorphic testing, novelty thresholds, observability and export rules. They are versioned and signed; see the Cartridges Pack documentation for details.

Commands

Use these commands in your LLM session to manage runners and cartridges:

attach runner <name>
attach cartridge <name>
detach runner <name>
detach cartridge <name>
show runners | show cartridges

You can attach multiple runners and cartridges simultaneously, but keep the total number of active cartridges to a reasonable limit (e.g. ≤ 7) to control cost and complexity. Cartridges are automatically deduplicated when multiple recommendations apply.

See the Master Runners Codex and Cartridges Pack for detailed descriptions of each runner and cartridge, including supported phases and configuration options.