haiku-assessment / n20/broader/prompts (Round 3)

Vanilla and down-skilled prompts for the broader probe: T3b (recalibrated voice rewrite), T4 JSON extraction, T5 changelog, T6 sort-and-filter, T7 NL→CLI, T8 copyedit.

• T3b_downskilled.md — 6.5 KB
• T3b_vanilla.md — 627 bytes
• T4_downskilled.md — 2.4 KB
• T4_vanilla.md — 230 bytes
• T5_downskilled.md — 2.1 KB
• T5_vanilla.md — 280 bytes
• T6_downskilled.md — 2.0 KB
• T6_vanilla.md — 450 bytes
• T7_downskilled.md — 2.0 KB
• T7_vanilla.md — 129 bytes
• T8_downskilled.md — 2.7 KB
• T8_vanilla.md — 394 bytes