--- name: description: | --- --- name: description: | <在 200 字符内说明用户意图、核心输入与交付物、适用场景、自然触发表达和一项重要边界。> --- # When the user wants to , follow this workflow. ## STEP 0: CHECK RESOURCES 1. ****: If not provided, ask the user to provide one or . 2. ****: Ask the user: "". 3. ****: If user specifies , do first. 4. Get from the input (e.g., duration, dimensions, format). ## STEP 1: ANALYZE INPUT Analyze the input to understand: - (e.g., mood, style, structure) - (e.g., content breakdown, sections) - (e.g., technical properties) Use this analysis throughout the workflow: - **Step 2**: Guide - **Step 3**: Determine ## STEP 2: GENERATE PLAN / SCRIPT Generate a complete that includes: ### (e.g., Characters, Themes, Sections) ```json { "": "", "": "", "": "" } ``` - : -- - : -- ### (e.g., Scenes, Layouts, Segments) ```json { "": "", "": "", "": "", "": "" } ``` **Timing / ordering rules**: - : e.g., segments must be continuous, no gaps - : e.g., each segment 3-15 seconds - : e.g., 7-10 seconds per segment -- fewer, longer segments produce more coherent results **Type / category rules**: - ``: - ``: **Ratio / balance rules**: - : e.g., ~60% type_1, ~40% type_2 by total duration - : e.g., never place two segments back-to-back ### Validate After generating the plan, validate it: - - - Fix all errors and re-validate until passed. ## STEP 3: GENERATE ASSETS Generate all in one batch: 1. **** (e.g., character images): Use each item's ``. : ``. 2. **** (e.g., scene images): Use each item's ``. : ``. Include ALL prompts in one task to minimize round-trips. **Pitfall**: Do NOT -- . ## STEP 4: CONFIRM WITH USER Present all generated assets to the user. Ask if any need adjustments. Regenerate as needed. ## STEP 5: MAIN PRODUCTION ### Step 5a: Prepare (batch) For each , prepare its : - - : `` -- **Batch**: Process ALL items in one call, then proceed to 5b. ### Step 5b: Generate (batch) Generate for all items: - : from Step 5a - : **Model selection**: . **Duration / size strategy**: : - -> - -> ### Step 5c: Adjust / Post-process Adjust every output to match its target specification: #### Case 1:

-> #### Case 2:

-> **Verification**: After adjusting all outputs, verify the total matches expectations. If drift exceeds , fix before proceeding. ### Step 5d: Generate (different technique) For , use because . **Sub-step 1**: Prepare inputs for this subset. **Sub-step 2**: Generate in one batch call. **Sub-step 3**: Post-process to match target specs. ## STEP 6: FINAL ASSEMBLY Assemble the final output: - : ALL produced outputs in order - : Original source material (e.g., audio track) - : Metadata (e.g., credits, annotations) ## STEP 7: PRESENT RESULT Show the final output to the user with a summary: - (e.g., source info) - (e.g., asset counts, techniques used) -