Character Portrait Scene Planner

character-portrait-scene-planner

Purpose

Translate a character-portrait intent (a single character in a hero / editorial / mood frame) into a structured scene plan that downstream quality / runtime / handoff skills validate. Distinct from product-hero-scene-planner (which targets manufactured products).

Quick start

• pin subject identity (real / stylized / archetype) + age + body type + role
• pin pose + expression + gaze
• pin wardrobe + props
• pin framing + camera + lens
• pin lighting + background + mood
• emit a structured plan + handoff to quality checkers

When to use

• intent is explicitly a character / person / creature portrait
• before any modeling or scene file is built
• when the consumer is editorial / brand / catalog / cinematic

When not to use

• product hero scenes (use product-hero-scene-planner)
• environment establishing shots (use environment-establishing-shot-planner)
• interior arch viz (use interior-architectural-scene-planner)
• multi-character ensemble (this skill is single-subject)

Trigger phrases

• "portrait of"
• "character looking at the camera"
• "headshot"
• "hero shot of [person / mascot / creature]"

Prerequisites / readiness

• brief from intent-to-3d-brief-writer is locked
• character identity + style is known
• target aspect + use case is known

Input schema

Required inputs

| Input | Why it is required | |---|---| | Subject identity | Determines proportions / rig / asset strategy | | Pose + expression + gaze | Drives camera and lighting decisions | | Framing intent | Close-up / medium / wide | | Aspect ratio | Required for composition rules |

Optional inputs

| Input | Use | |---|---| | Reference images | Calibrate proportions / style | | Wardrobe / prop list | Adds asset rows | | Existing rig / model | Reuse instead of re-build |

Assumptions to confirm

• Single subject only.
• Camera is locked unless animation called out.
• Style is consistent across subject + wardrobe + background.

Output schema

Primary output

A structured character portrait plan with sections: Subject, Pose / expression / gaze, Wardrobe + props, Camera / framing / lens, Lighting, Background, Mood, Style notes.

Secondary output

• asset list flagged for blender-asset-discovery-planner
• next-skill handoff list (composition / lighting / quality checkers)
• explicit Not Run / Not Produced labels

Evidence / caveat output

Runtime status: Not Run | Attempted | Produced | Verified | Failed | Blocked / Not Run
Artifact status: Not Run | Not Produced | Produced | Verified | Failed
Evidence used: 
Limitations: 

Required laws

• ../../laws/evidence-before-done.md
• ../../laws/non-blender-user-language.md
• ../../laws/no-arbitrary-python-interface.md
• ../../laws/official-runtime-only.md

Official runtime boundary

Planning skill — does not run Blender, model, or render. Output is a textual plan downstream quality and runtime skills validate against. If brief is unclear, hand off to intent-to-3d-brief-writer.

Operating procedure

1. Read brief; confirm subject identity + framing intent.
2. Fill subject section (age / body type / archetype / style).
3. Fill pose / expression / gaze (one decision per row).
4. Fill wardrobe + props.
5. Fill camera + framing + lens (per references/portrait-framing-rules.md).
6. Fill lighting plan (key + fill + rim, HDRI presence).
7. Fill background + mood.
8. List assets needed; flag for asset discovery.
9. Emit handoff to quality / lighting / composition checkers.

Decision tree

Subject is identifiable real person?
  → Add Warn for likeness rights + reference image needed
Pose unclear?
  → Hand off to intent-to-3d-brief-writer
Framing close-up but lens 18mm?
  → Warn (likely distortion); propose 50–85mm
Background detailed but framing tight close-up?
  → Warn (background mostly out of frame; simplify)

Playbooks

Playbook A: Hero close-up (16:9 web)

50–85mm equivalent; eyes on upper third; key + soft fill; rim for separation; simple gradient or studio HDRI background.

Playbook B: Editorial 9:16 mobile

24–35mm; subject head + upper body; environmental story background; mixed warm + cool lighting.

Playbook C: Stylized character mascot

Style-specific; flat or simplified background; key + bounce; consistent stylization across face + wardrobe.

Mode handling

Text-only mode

Pure plan; mark Artifact status: Not Run.

Runtime-ready mode

Plan plus runtime path attribution; do not upgrade to Verified without rendered evidence.

Blocked runtime mode

Plan only; flag what cannot be confirmed.

Validation checklist

• [ ] Subject section filled
• [ ] Pose / expression / gaze explicit
• [ ] Wardrobe + props listed
• [ ] Camera + framing + lens chosen
• [ ] Lighting plan present
• [ ] Background + mood explicit
• [ ] Style notes present
• [ ] Asset list flagged for discovery
• [ ] Handoff named

Pass / Warn / Fail rubric

| Verdict | Criteria | |---|---| | Pass | All sections filled; no "TBD"; framing rules respected; assets flagged. | | Warn | One section ambiguous (e.g. mood line missing); assets list partial. | | Fail | Subject identity unclear; framing impossible (15mm + tight close-up); no lighting plan. |

Failure handling

• Brief unclear → intent-to-3d-brief-writer.
• Style inconsistent → flag and ask user to lock style.
• Likeness rights unknown → Warn until confirmed.

Troubleshooting

| Problem | Response | |---|---| | Real person reference but no rights | Pause; require rights confirmation. | | Stylization clashes with wardrobe | Re-lock style early; cheaper than later. | | Lighting + background fight subject | Re-balance per lighting-quality-checker rules. |

Best practices

• Lock style early; revisions are cheap pre-modeling.
• Pair pose decisions with camera decisions in one row.
• Always flag assets for blender-asset-discovery-planner before committing modeling work.

Good examples

• "Subject: stylized 30s tech founder, neutral build. Pose: 3/4 toward camera, slight smile, gaze direct. Wardrobe: dark turtleneck. Camera: 85mm equivalent, eye-line upper third, 16:9. Lighting: soft key 250W + fill 50W + rim 100W; HDRI 0.7 studio. Background: gradient gray. Mood: confident editorial."

Bad examples

• "Person looking at camera." — no proportions, lighting, or framing.
• "Verified ready" — no rendered evidence.

User-facing response template

Subject: 
Pose / expression / gaze: 
Wardrobe + props: 
Camera / framing / lens:  |  |  | 
Lighting:  |  |  | 
Background: 
Mood + style notes: 

Assets to discover: 
Limitations: 
Next: blender-asset-discovery-planner / composition-quality-checker / lighting-quality-checker

Anti-patterns

• Performing modeling decisions inside this skill.
• Inventing likeness without rights confirmation.
• Skipping mood / style notes.
• Mixing multi-character or environment scope.

Cross-skill handoff

• Brief writer → ../intent-to-3d-brief-writer/SKILL.md
• Asset discovery → ../blender-asset-discovery-planner/SKILL.md
• Composition validation → ../composition-quality-checker/SKILL.md
• Lighting validation → ../lighting-quality-checker/SKILL.md
• Material validation → ../material-quality-checker/SKILL.md
• Quality readiness → ../blender-scene-quality-checker/SKILL.md

Non-goals

• Model the character.
• Animate the character.
• Render.
• Run Blender.

References

• references/portrait-framing-rules.md
• references/portrait-lighting-patterns.md
• references/style-consistency-rules.md
• ../../docs/skill-system.md

Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

• Author: ThanhNguyxnOrg
• Source: ThanhNguyxnOrg/blendops
• License: MIT
• Homepage: https://github.com/ThanhNguyxnOrg/blendops

Install and usage instructions live in the source repository linked above.