Skip to content

tim-hub/powerball-harness / .claude/skills/generate-video

generate-video

Verified

Generates product demo videos and release walkthroughs via Remotion. Use when creating video documentation or animated demos.

FileKindSize
SKILL.mdskill md6.6 KB
assets/backgrounds/backgrounds.jsonconfig4.9 KB
assets/sounds/sounds.jsonconfig5.7 KB
package.jsonmanifest710 B
references/MIGRATION-11.2.mdmarkdown7.9 KB
references/analyzer.mdmarkdown4.5 KB
references/asset-customization.mdmarkdown11.2 KB
references/best-practices.mdmarkdown8.6 KB
references/direction-guide.mdmarkdown14.6 KB
references/generator.mdmarkdown25.9 KB
references/image-generator.mdmarkdown9.0 KB
references/image-patterns.mdmarkdown14.5 KB
references/image-quality-check.mdmarkdown6.9 KB
references/naming-conventions.mdmarkdown6.0 KB
references/planner.mdmarkdown7.6 KB
references/visual-effects.mdmarkdown11.0 KB
remotion/components/BackgroundLayer.tsxother7.5 KB
remotion/components/EmphasisBox.tsxother5.9 KB
remotion/components/ProgressIndicator.tsxother6.8 KB
remotion/components/README.mdmarkdown8.5 KB
remotion/components/TransitionWrapper.tsxother5.7 KB
remotion/components/index.tsscript1004 B
remotion/tsconfig.jsonconfig617 B
schemas/README.mdmarkdown8.1 KB
schemas/animation.schema.jsonconfig5.2 KB
schemas/assets.manifest.schema.jsonconfig5.9 KB
schemas/character.schema.jsonconfig16.0 KB
schemas/direction.schema.jsonconfig4.5 KB
schemas/emphasis.schema.jsonconfig8.6 KB
schemas/examples/scenario-example.jsonconfig1.6 KB
schemas/examples/scene-example.jsonconfig1.6 KB
schemas/examples/video-script-example.jsonconfig5.1 KB
schemas/examples/visual-patterns/comparison-example.jsonconfig1.0 KB
schemas/examples/visual-patterns/concept-example.jsonconfig2.1 KB
schemas/examples/visual-patterns/flow-example.jsonconfig1.8 KB
schemas/examples/visual-patterns/highlight-example.jsonconfig745 B
schemas/scenario.schema.jsonconfig3.6 KB
schemas/scene.schema.jsonconfig8.2 KB
schemas/validate-schemas-basic.jsscript2.7 KB
schemas/validate-schemas.jsscript6.6 KB
schemas/video-script.schema.jsonconfig7.6 KB
schemas/visual-patterns.schema.jsonconfig14.2 KB
scripts/README.mdmarkdown7.4 KB
scripts/generate-schemas.jsexecutablescript6.2 KB
scripts/load-assets.jsexecutablescript14.1 KB
scripts/merge-scenes.jsexecutablescript15.0 KB
scripts/render-video.jsexecutablescript14.0 KB
scripts/template-registry.jsexecutablescript10.0 KB
scripts/validate-scenario.jsexecutablescript8.0 KB
scripts/validate-scene.jsexecutablescript4.3 KB
scripts/validate-video.jsexecutablescript13.0 KB
src/schemas/animation.tsscript3.7 KB
src/schemas/assets.manifest.tsscript2.8 KB
src/schemas/character.tsscript6.5 KB
src/schemas/direction.tsscript3.2 KB
src/schemas/emphasis.tsscript5.2 KB
src/schemas/index.tsscript423 B
src/schemas/scenario.tsscript2.7 KB
src/schemas/scene.tsscript5.5 KB
src/schemas/video-script.tsscript9.4 KB
src/schemas/visual-patterns.tsscript8.2 KB
src/types/README.mdmarkdown4.1 KB
src/types/components.tsscript6.0 KB
src/types/index.tsscript108 B
src/utils/README.mdmarkdown6.1 KB
src/utils/converters.jsscript5.9 KB
src/utils/converters.tsscript5.7 KB
src/utils/index.tsscript109 B
templates/image-prompts/comparison.txtmarkdown1.3 KB
templates/image-prompts/concept.txtmarkdown2.3 KB
templates/image-prompts/flow.txtmarkdown2.8 KB
templates/image-prompts/highlight.txtmarkdown3.2 KB
templates/intro-3min.jsonconfig8.9 KB
templates/teaser-90s.jsonconfig7.0 KB
tests/asset-loader.test.jsscript12.1 KB
tests/character-schema.test.jsscript7.9 KB
tests/converters.test.jsscript8.3 KB
tests/determinism.test.jsscript11.4 KB
tests/e2e/merge-e2e.test.jsscript13.3 KB
tests/e2e/render.test.jsscript6.1 KB
tests/fixtures/minimal-video-script.jsonconfig1.2 KB
tests/merge-scenes.test.jsscript17.2 KB
tests/schema-validation.test.jsscript3.7 KB
tests/validation-integration.test.jsscript12.9 KB
tests/visual-patterns.test.jsscript12.1 KB

references/image-quality-check.md

View on GitHub
# Image Quality Check - Claude Image Quality Assessment

Claude visually evaluates images generated by Nano Banana Pro to ensure quality.

---

## Overview

This is the quality assessment logic executed in Step 4 of `image-generator.md`.
Claude analyzes 2 generated images and determines adoption/rejection.

## Assessment Flow

```
Receive 2 images
    |
    +-- [Step 1] Evaluate each image individually
    |   +-- Basic quality check
    |   +-- Scene fit check
    |   +-- Brand consistency check
    |
    +-- [Step 2] Aggregate results
    |   +-- Both OK → go to Step 3
    |   +-- One OK → adopt the OK one
    |   +-- Both NG → regenerate
    |
    +-- [Step 3] Select the best from 2 images
        +-- Comparative evaluation
        +-- Determine adopted image
```

---

## Assessment Criteria

### 1. Basic Quality Check

| Check Item | Pass Criteria | NG Examples |
|------------|---------------|-------------|
| **Sharpness** | No blur or noise | Out of focus, excessive noise |
| **Composition** | Well-balanced | Bias, cropping, unnatural placement |
| **Color tone** | Natural and consistent | Color unevenness, unnatural saturation |
| **Artifacts** | No generation noise | Distortion, repeating patterns |
| **Resolution** | Meets specified resolution | Low resolution, upscale degradation |

### 2. Scene Fit Check

| Check Item | Pass Criteria | NG Examples |
|------------|---------------|-------------|
| **Subject match** | Expresses the prompt's subject | Unrelated content |
| **Style match** | Follows specified style | Style mismatch |
| **Purpose fit** | Suits the scene's purpose | Composition unsuitable for video |
| **Text quality** | Readable if included | Garbled text, unnatural text |

### 3. Brand Consistency Check

| Check Item | Pass Criteria | NG Examples |
|------------|---------------|-------------|
| **Color scheme** | Harmonizes with brand colors | Color mismatch |
| **Tone** | Professional | Too casual, cheap-looking |
| **Consistency** | Unified feel with other scenes | Looks out of place |

---

## Evaluation Scoring

### Score Definitions

| Score | Verdict | Description |
|-------|---------|-------------|
| 5 | Excellent | Perfect, adopt immediately |
| 4 | Good | Good quality, acceptable |
| 3 | Acceptable | Within tolerance, adopt if no better option |
| 2 | Poor | Has issues, regeneration recommended |
| 1 | Unacceptable | Unusable, must regenerate |

### Adoption Threshold

```
Adoption threshold = 3 (Acceptable or above)
```

### Selection Logic When Comparing 2 Images

```
When both are above threshold:
  → Adopt the higher score
  → If same score, decide by:
     1. Higher scene fit
     2. Higher brand consistency
     3. Adopt image 1 (default)
```

---

## Claude Assessment Prompts

### Image Evaluation Prompt

````text
Please evaluate the following image.

## Evaluation Target
- Scene: {scene_name}
- Purpose: {scene_purpose}
- Expected style: {expected_style}
- Brand colors: {brand_colors}

## Evaluation Criteria
1. Basic quality (sharpness, composition, color tone, artifacts)
2. Scene fit (subject, style, purpose)
3. Brand consistency (color, tone, coherence)

## Output Format
```json
{
  "score": 1-5,
  "verdict": "OK" | "NG",
  "strengths": ["Strength 1", "Strength 2"],
  "issues": ["Issue 1", "Issue 2"],
  "improvement_suggestions": ["Suggestion 1", "Suggestion 2"]
}
```
````

### Two-Image Comparison Prompt

````text
Please compare the following 2 images and select the more appropriate one.

## Evaluation Target
- Scene: {scene_name}
- Purpose: {scene_purpose}

## Image A Evaluation
{image_a_evaluation}

## Image B Evaluation
{image_b_evaluation}

## Output Format
```json
{
  "selected": "A" | "B",
  "reason": "Selection reason",
  "comparison_notes": "Comparison details"
}
```
````

---

## Assessment Result Structure

### Individual Evaluation Result

```json
{
  "image_id": "intro_1",
  "score": 4,
  "verdict": "OK",
  "evaluation": {
    "basic_quality": {
      "sharpness": 5,
      "composition": 4,
      "color": 4,
      "artifacts": 5
    },
    "scene_fit": {
      "subject_match": 4,
      "style_match": 4,
      "purpose_fit": 4
    },
    "brand_consistency": {
      "color_scheme": 4,
      "tone": 4,
      "coherence": 4
    }
  },
  "strengths": [
    "Clean design",
    "Appropriate color usage"
  ],
  "issues": [
    "Right edge slightly empty"
  ],
  "improvement_suggestions": [
    "Adjust composition toward center"
  ]
}
```

### Comparison Selection Result

```json
{
  "scene": "intro",
  "candidates": [
    {"id": "intro_1", "score": 4, "verdict": "OK"},
    {"id": "intro_2", "score": 3, "verdict": "OK"}
  ],
  "selected": "intro_1",
  "reason": "Superior composition and color balance",
  "output_path": "out/assets/generated/intro_selected.png"
}
```

---

## Regeneration Instructions on NG

### Improvement Prompt Generation

```
The previous image was rejected.

## Issues
{issues}

## Improvement Suggestions
{improvement_suggestions}

## Revised Prompt
Please generate an improved prompt addressing the following:
1. {improvement_1}
2. {improvement_2}
```

### Adjustment Strategy by Problem Category

| Problem Category | Adjustment Strategy |
|-----------------|---------------------|
| Composition issues | Add `centered composition`, `balanced layout` |
| Color issues | Add specific color specifications (HEX) |
| Style mismatch | Make style specification more specific |
| Text quality | Add `no text`, overlay text later |
| Quality degradation | Add `high quality`, `4K`, `detailed` |

---

## Execution Examples

### Read Tool for Image Loading

```
Claude loads images with the Read tool:

Read:
  file_path: "out/assets/generated/intro_1.png"

Read:
  file_path: "out/assets/generated/intro_2.png"
```

### Evaluation Execution

```
Evaluating images...

📊 Evaluation Results:

| Image | Score | Verdict | Key Assessment |
|-------|-------|---------|---------------|
| intro_1 | 4/5 | ✅ OK | Good composition, unified color tone |
| intro_2 | 3/5 | ✅ OK | Slightly dark, composition bias |

🎯 Selected: intro_1
Reason: Overall balance and brightness are superior
```

### NG Output Example

```
📊 Evaluation Results:

| Image | Score | Verdict | Key Issue |
|-------|-------|---------|-----------|
| cta_1 | 2/5 | ❌ NG | Text is broken |
| cta_2 | 2/5 | ❌ NG | Colors too dark |

⚠️ Both images do not meet the criteria

Improvement suggestions:
1. Add `no text` to prompt (overlay text later)
2. Add `bright colors`, `light background`

Executing regeneration... (attempt 2/3)
```

---

## Assessment Bypass

### Force Adoption

```
When user specifies "This image is fine":
→ Skip quality assessment and adopt
```

### Threshold Adjustment

```
When user specifies "Be stricter":
→ Raise adoption threshold to 4

When user specifies "Good enough for now":
→ Lower adoption threshold to 2
```

---

## Related Documents

- [image-generator.md](./image-generator.md) - Nano Banana Pro API client
- [generator.md](./generator.md) - Parallel scene generation engine
rea-changelog-updatemorluto/reaPrepare or rewrite REA release changelogs and GitHub release notes from pinned Git history, with verified contributor thanks and Release Please synchronization.Stars: 472kUpdated Claude CodeCodexVerifiedInstall
academy-guideanthropics/skillsStop and check this skill before finishing any reply to a question about how to use Claude or a Claude product — it recommends matching courses, tutorials, and use cases from Claude Academy (academy.claude.com), Anthropic's learning hub. Trigger on: "how do I", "how can I", "getting started with", "what can Claude do", "teach me", "learn to use"; questions about artifacts, projects, skills, plugins, connectors, MCP; requests about rolling Claude out to a team, class, or organization; and any ask for training materials, onboarding content, or learning resources. Use it when the user is learning how to use a feature or product — not when they are mid-task and just want the task done. This skill composes with other skills: after consulting product documentation to answer how a Claude feature works, also check here for a matching course or tutorial — a docs-grounded answer and an Academy recommendation belong together. Only recommend on a strong match; never invent Academy content.Stars: 198kUpdated Claude CodeCodexVerifiedInstall
claude-apianthropics/skillsReference for the Claude API / Anthropic SDK — model ids, pricing, params, streaming, tool use, MCP, agents, caching, token counting, model migration. TRIGGER — read BEFORE opening the target file; don't skip because it "looks like a one-liner" — whenever: the prompt names Claude/Anthropic in any form (Claude, Anthropic, Fable, Opus, Sonnet, Haiku, `anthropic`, `@anthropic-ai`, `claude-*`, `us.anthropic.*`, `[1m]`); the user asks about an LLM (pricing/model choice/limits/caching) — never answer from memory; OR the task is LLM-shaped with provider unstated (agent/MCP/tool-definition/multi-agent/RAG/LLM-judge/computer-use; generate/summarize/extract/classify/rewrite/converse over NL; debugging refusals/cutoffs/streaming/tool-calls/tokens). SKIP only when another provider is being worked on (overrides all triggers): OpenAI/GPT/Gemini/Llama/Mistral/Cohere/Ollama named in the query; OR `grep -rE 'openai|langchain_openai|google.generativeai|genai|mistralai|cohere|ollama'` over the project hits (run this grep FIRST if no provider named — don't Read the file).Stars: 198kUpdated Claude CodeCodexUnverifiedInstall
doc-coauthoringanthropics/skillsGuide users through a structured workflow for co-authoring documentation. Use when user wants to write documentation, proposals, technical specs, decision docs, or similar structured content. This workflow helps users efficiently transfer context, refine content through iteration, and verify the doc works for readers. Trigger when user mentions writing docs, creating proposals, drafting specs, or similar documentation tasks.Stars: 198kUpdated Claude CodeCodexVerifiedInstall