| name | generate-voice-lines |
| description | Generate voice lines aligned with the final storyboard panel set. |
Generate Voice Lines
Purpose
Generate structured voice-over or spoken line output aligned with the final storyboard panel set. This is the final skill in the script-to-storyboard workflow.
Use This Skill When
storyboard.panel_set exists and is current.
- The storyboard workflow needs voice output aligned to final panels.
- Storyboard detail changed and prior voice lines are no longer trustworthy.
Do Not Use This Skill When
- Final storyboard panels are missing.
- The caller only needs storyboard structure without voice output.
- Upstream panel alignment is unresolved.
Inputs
storyboard.panel_set
- Story text for narrative grounding
- Voice generation context derived from final storyboard output
Outputs
- Structured voice line payload.
- Persistable payload for the
voice.lines artifact.
- UI summary describing line count and clip coverage.
Artifact Contract
- Reads:
storyboard.panel_set
- Writes:
voice.lines
- Does not replace upstream storyboard artifacts, but depends on them being valid.
Execution Rules
- Must run after
refine-storyboard-detail.
- Preserve clip and panel alignment from the final panel set.
- Do not invent voice lines disconnected from the storyboard structure.
- Fail explicitly if alignment or schema validation breaks.
Failure Conditions
storyboard.panel_set is missing or invalid.
- Voice lines cannot be aligned to final panels.
- Model output fails schema validation.
- The result leaves clips without required voice coverage.
Quality Bar
- Voice lines must stay grounded in the final board and source story.
- Output should be structured, concise, and ready for downstream preview or generation.
- The skill must fail rather than emit ambiguous orphan lines.