Instructions to compress biolographical/character summary and paragraphs to an LLM token-efficient format.

SemanticDenseBlock.md

Semantic Dense Block (SDB): Single‑paragraph token-efficient compression format removing articles/linking verbs while preserving vivid descriptors.

Structure: One block, `Header:Value.` syntax, capitalized headers, periods after sections, commas/slashes for lists, no line breaks.

ID Header: `Name[Age,Gender,Height,Species,Role].`

Header Map: 
Appearance=Look, Clothing=Wear, Personality=Mind, Skills=Skills, Speech=Voice.

Rules: Drop articles/linking verbs, keep flavorful adjectives, group related details (e.g., `Look:White hair + clip.`).

Instruction: Convert character data into one paragraph using short headers, no JSON/XML, remove filler words, preserve aesthetic adjectives.

Example Input: "She wears a heavy combat jacket that is dark green. She is usually very cynical and doesn't trust people easily, but she is loyal to her squad."

SDB Output: 
```text
Wear:Heavy dark green combat jacket. Mind:Cynical, slow-to-trust, loyal to squad.
```