lighting-quality-checker

Purpose

Run a focused pass/warn/fail check on the scene's lighting setup per light — covering intensity ranges, shadow softness, color temperature, HDRI / explicit-light balance, and mood match — without expanding into material, composition, or lookdev planning concerns.

Quick start

• enumerate every explicit light and HDRI in the scene plan or measured scene
• check intensity ranges per light
• check shadow softness consistency
• check color temperature consistency
• check HDRI presence + balance with explicit lights
• check mood match against the brief
• emit per-light verdicts (Pass / Warn / Fail) with remediation hints only + a top-line status

When to use

• before render
• after a lighting change request
• as a sub-check during full quality review
• when the rendered preview looks "off" but the issue is unclear

When not to use

• upstream lookdev and lighting planning → use ../lighting-material-plan/SKILL.md (use lighting-material-plan)
• material validation → use ../material-quality-checker/SKILL.md (use material-quality-checker)
• composition / camera validation → use ../composition-quality-checker/SKILL.md (use composition-quality-checker)
• broad pass/warn/fail scene readiness → use ../scene-readiness/SKILL.md (use scene-readiness)
• classifying render or export artifact evidence → use ../render-export-evidence/SKILL.md (use render-export-evidence)
• to fix or edit lights (this skill validates and gives remediation hints only; it never edits) → use ../scene-readiness/SKILL.md (this skill validates and gives remediation hints only; it never edits)

Trigger phrases

• "is the lighting okay"
• "check shadow softness"
• "lighting feels off"
• "is the HDRI balanced"

Prerequisites / readiness

• light list known (from lighting-material-plan scene plan or measured scene snapshot)
• consumer target known (render or export target)
• mood / look brief known (from intent-to-3d-brief-writer)

Input schema

Required inputs

Optional inputs

Assumptions to confirm

• Intensity units (Watts / Lumens / Sun strength) are consistent across the scene.
• Color temperature is in Kelvin where applicable; mixed Kelvin and RGB color modes need explicit annotation.
• HDRI is real-world calibrated (typical 0.5–2.0 strength), not arbitrarily boosted.

Output schema

Primary output

A per-light verdict table: name, type, intensity, color, shadow softness, verdict (Pass / Warn / Fail), and remediation hints (hints only, never direct edits or execution).

Secondary output

• top-line status (Pass / Warn / Fail) computed from per-light rows
• mood / look match: Yes / No / Unknown
• worst-offender highlighting

Evidence / caveat output

Validation verdicts (Pass / Warn / Fail) are separate from runtime and artifact truth states. Truth states are downgrade-only (never upgrade), and Artifact status defaults to Not Produced unless concrete output evidence exists (classified via render-export-evidence).

Runtime status: Not Run | Attempted | Produced | Verified | Failed
Artifact status: Not Produced | Produced | Verified | Failed
Evidence used: <links, paths, logs, or "none">
Limitations: <known gaps>

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

This skill validates plans or measured scene snapshots only. It never modifies lights, runs Blender, or renders. Use measured scene description (plan output, runtime tool snapshot) — never invent values.

For runtime path attribution, see ../../docs/runtime-stack-strategy.md.

Operating procedure

1. Read the light list from the lighting-material-plan plan or measured scene snapshot.
2. For each light, evaluate intensity range (see references/intensity-ranges.md).
3. Evaluate shadow softness consistency (see references/shadow-softness-rules.md).
4. Evaluate color temperature consistency.
5. Evaluate HDRI strength + explicit-light balance (see references/hdri-balance-rules.md).
6. Cross-check overall setup with the mood brief from intent-to-3d-brief-writer.
7. Assign per-light verdict (Pass / Warn / Fail) and provide brief remediation hints only (never edit lights).
8. Compute top-line status, highlight worst offenders, and record downgrade-only truth states (Artifact status: Not Produced by default unless concrete output evidence exists).
9. Hand off to scene-readiness (broad review), lighting-material-plan (re-plan), material-quality-checker (materials), composition-quality-checker (framing), or render-export-evidence (artifact evidence classification).

Decision tree

Scene has only HDRI, no explicit light when brief requires strong key separation?
  → Warn (HDRI/explicit-light balance flag; verify intent)
HDRI strength overpowers explicit lights?
  → Warn (balance issue); add remediation hint to rebalance HDRI vs key light
Multiple explicit lights stacked above target intensity range?
  → Warn or Fail depending on severity; add remediation hint to reduce intensities
Shadows hard when mood brief specifies soft lighting (or vice versa)?
  → Fail (shadow softness + mood mismatch)
Inconsistent or unannotated color temperatures across lights?
  → Warn (color temperature consistency flag)

Playbooks

Playbook A: Product hero with HDRI + 3-point lighting

HDRI ~0.7–1.2; key area light ~50–500W; fill ~10–20% of key; rim for separation. Verdicts per light + balance check.

Playbook B: Cinematic moody scene

Single key + practical lights; HDRI ~0.2 (mostly ambient fill); shadows hard or soft per brief; verify mood match.

Playbook C: Outdoor sunlight

Sun strength ~1–5; HDRI environment map for ambient; verify shadow direction matches sun direction; verify color temperature (~5500–6500K daylight).

Mode handling

Text-only mode

Verdicts (Pass / Warn / Fail) are based on the plan or spec only. Set Runtime status: Not Run and Artifact status: Not Produced.

Runtime-ready mode

Verdicts are based on a measured scene snapshot; cite the inspection source used. Truth states are downgrade-only — never upgrade Runtime status or Artifact status inside this skill, and leave Artifact status: Not Produced unless concrete output evidence already exists and is classified via render-export-evidence.

Blocked runtime mode

Skip rows that depend on missing live measurement, mark limitations explicitly, and downgrade truth states (Not Run / Not Produced / Failed as applicable).

Validation checklist

• Every light listed
• Intensity column filled
• Color column filled (Kelvin or RGB)
• Shadow softness column filled
• HDRI presence noted (or "none")
• Per-light verdict assigned
• Mood match assessed
• Top-line status derived from rows
• Honest evidence label

Pass / Warn / Fail rubric

Failure handling

• Missing light data → Fail with note.
• Mood unclear → mark Mood match Unknown; do not Pass.
• User pushes for "Pass" without checks → restate requirement.

Troubleshooting

Best practices

• Pair every Fail with the rule that triggered it.
• Cross-check intensity sums against the scene's overall exposure target.
• Always evaluate HDRI + explicit lights as a system, not separately.
• Keep this skill narrow — do not creep into material or composition.

Good examples

• "HDRI: studio.exr @ 1.0 (Pass); Key area light 250W, 5500K, 0.6m soft (Pass); Fill area 50W, 5500K, 0.8m soft (Pass); Rim area 80W, 5000K, 0.4m soft (Pass). Mood match: Yes (soft hero). Top-line: Pass. Runtime status: Not Run | Artifact status: Not Produced."

Bad examples

• "Lighting looks fine." — missing per-light rows, mood match, and top-line status.
• Upgrading Artifact status to Verified or Produced inside a lighting check without concrete output evidence classified by render-export-evidence.
• Editing light power or running Blender scripts instead of giving remediation hints.

User-facing response template

Consumer target: <target>
Mood / look: <one-line summary>

| Light | Type | Intensity | Color | Shadow softness | Verdict | Remediation hint |
|---|---|---|---|---|---|---|
| <name> | <type / HDRI> | <value> | <color temp / value> | <softness> | <Pass / Warn / Fail> | <hint only, no edits> |

HDRI balance: <name + strength + balance note, or "none">
Mood match: Yes / No / Unknown
Top-line: <Pass / Warn / Fail>
Worst offenders: <list or "none">

Runtime status: Not Run | Attempted | Produced | Verified | Failed
Artifact status: Not Produced | Produced | Verified | Failed
Evidence used: <links, paths, logs, or "none">
Limitations: <known gaps>
Next: scene-readiness | lighting-material-plan | material-quality-checker | composition-quality-checker | render-export-evidence

Anti-patterns

• Editing lights or generating full lighting recipes inside this skill (use lighting-material-plan).
• Inventing light intensity, color, or softness values.
• Upgrading Runtime or Artifact truth states (truth states are downgrade-only; artifacts stay Not Produced unless concrete evidence exists).
• Mixing material checks (material-quality-checker), camera/framing checks (composition-quality-checker), or artifact evidence classification (render-export-evidence) into lighting validation.

Cross-skill handoff

• [progression] Broad readiness review → ../scene-readiness/SKILL.md
• [upstream_reference] Material validation → ../material-quality-checker/SKILL.md
• [progression] Composition validation → ../composition-quality-checker/SKILL.md
• [retry] Lighting re-plan → ../lighting-material-plan/SKILL.md
• [progression] Render evidence → ../render-export-evidence/SKILL.md

Non-goals

• Edit lights.
• Generate light setup recipes.
• Replace scene-readiness.
• Run Blender.

References

• references/intensity-ranges.md
• references/shadow-softness-rules.md
• references/hdri-balance-rules.md
• ../../laws/evidence-before-done.md
• ../../docs/skill-system.md