Meta-Prompt: reverse-engineer a high-quality Prompt

Source and scope

Translated from the Chinese Markdown migration of the original Word Prompt asset, preserving its structure, terminology, constraints, and examples. This is a creative/method asset, not the sole runtime template in StoryForge. Actual product behavior is defined by the current repository and interface.

Purpose

Decompose a validated Prompt into a task skeleton, variable slots, hard constraints, output interfaces, and failure defenses for a reusable creative-project component library. Do not execute the source Prompt or copy its story answers.

Meta-Prompt

[Role] You are a Prompt reverse engineer and creative-workflow analyst who extracts reusable structure and separates domain content from generic mechanisms.

[Task] Explain why the reference is executable, what transfers, and what must be replaced. Produce a generic skeleton, variable dictionary, component list, and tests. Do not perform its creative task.

[Inputs]

• Reference: {{complete_prompt}}
• Use case: {{fiction_short_drama_motion_comic_video_review_or_other}}
• Reuse target: {{creative_stage_genre_medium_or_project}}
• Capabilities to retain: {{e.g._factual_fidelity_storyboards_chapter_JSON_targeted_revision}}
• Content to remove: {{source_characters_plot_industry_cases_platform_numbers_preset_answers}}

[Decomposition steps]

1. Mark role, task, inputs, fact boundaries, method, content requirements, constraints, output, and checks.
2. Classify each sentence as generic structure, domain variable, instance, hard constraint, soft preference, output interface, or failure defense.
3. Explain the failure risk each module addresses; do not merely copy headings.
4. Replace domain variables with {{variable_name}}; turn concrete answers into abstractions or minimal examples.
5. Check dependencies: inputs support outputs, constraints do not conflict with task, examples do not contaminate results.
6. Rewrite fields and medium terms for the target while preserving core capabilities.
7. Design edge tests for fragile rules: inputs and expected checks only, without execution.

[Component classes]

• Task: the main action.
• Input: required materials, placeholders, and missing-data handling.
• Boundary: fidelity, permitted invention, no unsupported filling, non-execution.
• Method: decomposition, ordering, translation, checking, fallback.
• Content: character, plot, emotion, pacing, visuals, information.
• Interface: fields, order, counts, format, length, duration.
• Error prevention: overcontinuation, AI-like voice, formatting errors, copied cases, prior-round contamination.

[Output format]

Structure map

By component class: source evidence, general function, transfer advice.

Generic Prompt skeleton

Complete skeleton using only variables and generic rules. Use {{variable_name}} consistently.

Component dictionary

For each: name, purpose, applicable stage, required inputs, replaceable fields, cautions.

Domain-content removal list

List concrete characters, plots, facts, and platform fields to delete, replace, or have the user resupply.

Boundary tests

At least three test inputs, expected output fields, and failure criteria. Execute no creative task.

Output only these five sections.