Story Poet: AI-Free Poetry

Problem

When asked to turn a story into classical Chinese poetry, models often produce vague imagery, slogan-like endings, inaccurate emotional intensity, and wrong gendered self-reference. The hard part is not merely rhyming; it is preserving concrete actions, awkwardness, contrast, and causality from the source scene while respecting meter constraints for forms such as Quatrain, Regulated Verse, and Jiang Cheng Zi.

Workflow

The skill runs a three-stage pipeline:

• Input routing: detects whether the user provides a novel title, a story scene, or an existing draft, then enters source extraction, poetry drafting, or polishing.
• Scene extraction: for novel titles, it prioritizes public free TXT sources, validates file size and Chinese prose content, and extracts 3–5 high-stakes, conflict, monologue, or visual passages.
• Metered drafting: it infers user gender, tone, and form preference, then drafts poetry with concrete details, gender checks, intensity checks, and layered humor techniques.
• AI-flavor removal: it audits clichéd imagery, overuse of weak particles, excessive sentiment, and slogan endings, then revises line by line with concrete actions and stance.

Boundaries

Use it for public novels, life anecdotes, or existing drafts. It does not support paid content, DRM-protected text, or copyright circumvention. Output is derivative work, should credit the source novel, and should avoid political, sexual, or violent content.