Prose

To override quantization, activate a readable-prose mode that bypasses default token compression, forcing quantization-aware outputs to render as clear text. This improves clarity for explanations, onboarding, and handoffs by prioritizing readability over token density.

Quantization compresses outputs to maximize token density, which can degrade readability for human-facing text. Overcoming compressed outputs requires a cross-skill override that forces quantization-aware systems to render explanations in standard, readable prose.

Enable readable prose mode by issuing the slash command '/cellm:prose'. This applies a cross-skill override across all quantization-aware skills, and you can optionally specify level or TTL modifiers to control the duration and intensity of the override.

Yes, the prose override remains controllable via slash commands with TTL and level modifiers. It implements a filesystem-backed flag with safety checks, enforcing strict access through a safeWriteFlag contract to manage activation per skill.

Use a readable prose override for safety-critical explanations, multilingual communications, and documentation handoffs. It is necessary whenever quantization-aware token compression degrades the clarity required for human-readable explanations and onboarding materials.

Yes, the prose override applies across all quantization-aware skills. It uses a filesystem-backed flag with safety checks and enforces strict access through the safeWriteFlag contract, ensuring the readable-prose mode functions universally across compatible systems.