wicked-garden 12.27.0

package/skills/product/analyze/refs/sentiment-patterns.md

@@ -0,0 +1,180 @@
+---
+phase_relevance: ["clarify", "design", "review"]
+archetype_relevance: ["*"]
+---
+# Sentiment Analysis Patterns
+
+Common patterns and indicators for sentiment classification.
+
+## Language Patterns
+
+### Strong Positive
+- **Delight**: "I love this", "This is amazing", "Perfect solution"
+- **Gratitude**: "Thank you so much", "You saved me hours"
+- **Advocacy**: "I recommend this to everyone", "Best tool I've used"
+- **Surprise**: "Wow!", "I didn't expect this to work so well"
+
+Example:
+```
+"This feature is absolutely brilliant! Saved me hours of work.
+I'm recommending it to my whole team."
+
+→ Sentiment: POSITIVE (STRONG)
+```
+
+### Moderate Positive
+- **Satisfaction**: "This works well", "Good solution", "Helpful feature"
+- **Improvement**: "Better than before", "Nice upgrade"
+- **Acknowledgment**: "Appreciate the update", "Thanks for fixing"
+
+Example:
+```
+"The new dashboard is much better. Makes finding reports easier."
+
+→ Sentiment: POSITIVE (MODERATE)
+```
+
+### Mild Positive
+- **Acceptance**: "It's fine", "Works as expected", "Does the job"
+- **Neutral praise**: "Nice to have", "Useful in some cases"
+
+Example:
+```
+"Export feature is decent. Gets the job done."
+
+→ Sentiment: POSITIVE (MILD)
+```
+
+### Strong Negative
+- **Frustration**: "This is terrible", "Completely broken", "Unusable"
+- **Anger**: "I'm furious", "This is unacceptable"
+- **Desperation**: "Please fix this ASAP", "About to cancel"
+- **Churn signals**: "Looking at alternatives", "Switching to competitor"
+
+Example:
+```
+"This is COMPLETELY BROKEN! I've lost hours of work.
+If this isn't fixed soon, we're switching to [Competitor]."
+
+→ Sentiment: NEGATIVE (STRONG)
+→ Urgency: IMMEDIATE (churn risk)
+```
+
+### Moderate Negative
+- **Complaints**: "This is frustrating", "Doesn't work as expected"
+- **Confusion**: "I don't understand how to...", "This is confusing"
+- **Slowness**: "Too slow", "Takes forever to load"
+
+Example:
+```
+"The upload process is really slow and confusing.
+Not sure what format is expected."
+
+→ Sentiment: NEGATIVE (MODERATE)
+```
+
+### Mild Negative
+- **Minor issues**: "Small bug", "Could be better"
+- **Suggestions**: "Would be nice if...", "Consider improving"
+
+Example:
+```
+"Minor issue with date formatting. Not a big deal but could be clearer."
+
+→ Sentiment: NEGATIVE (MILD)
+```
+
+### Mixed Sentiment
+- **Qualified praise**: "I like X but Y is broken"
+- **Tradeoffs**: "Good for A, bad for B"
+- **Context-dependent**: "Works on desktop, broken on mobile"
+
+Example:
+```
+"Love the new features, but the mobile experience is terrible.
+Desktop version is great though."
+
+→ Sentiment: MIXED (positive desktop, negative mobile)
+```
+
+## Emotional Indicators
+
+### Intensity Markers
+- **Emphasis**: !!!, CAPS, bold, multiple adjectives
+- **Repetition**: "very very slow", "again and again"
+- **Profanity**: (indicates strong frustration)
+- **Time expressions**: "always", "never", "constantly"
+
+### Urgency Signals
+- "ASAP", "urgent", "critical"
+- "can't work", "blocking", "showstopper"
+- "about to cancel", "looking at alternatives"
+- Time constraints: "need by Friday", "deadline approaching"
+
+### Satisfaction Signals
+- Emoticons: :), :D, ❤️
+- Exclamation marks: "Great!"
+- Gratitude: "thank you", "appreciate"
+- Recommendation: "telling everyone about this"
+
+## Context Clues
+
+### Business Impact
+- Revenue mentions: "costing us money", "losing sales"
+- Productivity: "wasting hours", "manual workaround"
+- Scale: "affects entire team", "all our customers"
+
+### User Sophistication
+- Power user: Detailed technical feedback, feature requests
+- New user: Onboarding issues, basic confusion
+- Admin: Team management, permissions, configuration
+
+### Segment Patterns
+- Enterprise: Compliance, security, scale concerns
+- SMB: Price sensitivity, ease of use, support quality
+- Startup: Speed, integrations, modern features
+
+## Edge Cases
+
+### Sarcasm (Hard to Detect)
+```
+"Oh great, another bug. Just what I needed today."
+
+→ Likely NEGATIVE despite "great"
+→ Context: "another", "just what I needed" (sarcastic)
+```
+
+### Feature Requests (Often Neutral)
+```
+"Would love to see export to PDF feature."
+
+→ Sentiment: NEUTRAL (request, not complaint)
+→ But note "would love" shows positive attitude
+```
+
+### Bug Reports (Context-Dependent)
+```
+# Frustrated
+"This bug is driving me crazy! Please fix!"
+→ NEGATIVE (STRONG)
+
+# Helpful
+"Found a small bug in the export. Steps to reproduce: ..."
+→ NEUTRAL or POSITIVE (MILD) - constructive, helpful tone
+```
+
+## Analysis Tips
+
+1. **Read full context**: Don't just keyword match
+2. **Consider tone**: "I guess it works" vs "It works great!"
+3. **Watch for qualifiers**: "but", "however", "although"
+4. **Note segment**: Enterprise expectations differ from startup
+5. **Check recency**: Old complaint might be fixed
+6. **Look for patterns**: Multiple people, same issue = signal
+
+## False Positives to Avoid
+
+- **Quoting others**: "Some people say it's slow" ≠ user thinks it's slow
+- **Hypotheticals**: "If this breaks, it would be terrible" ≠ it's broken
+- **Questions**: "Is this slow for everyone?" ≠ complaint
+- **Feature requests**: "Add dark mode" ≠ current version is bad

package/skills/product/imagery/SKILL.md

@@ -0,0 +1,97 @@
+---
+name: imagery
+description: |
+  Visual asset lifecycle management — analysis, generation, modification, and review.
+  Three sub-skills cover the full creative pipeline with provider abstraction.
+
+  Use when: generating, editing, analyzing, or reviewing an image — routes to
+  the appropriate sub-skill (create / alter / review).
+phase_relevance: ["clarify", "design", "review"]
+archetype_relevance: ["*"]
+---
+
+# Imagery: Visual Asset Lifecycle
+
+Manage the full lifecycle of visual assets through three specialized sub-skills. Each sub-skill handles a distinct phase of creative work, from analysis through generation to iterative refinement.
+
+## Sub-Skills
+
+| Sub-Skill | Purpose | Provider Required |
+|-----------|---------|-------------------|
+| [**review**](review/SKILL.md) | Image analysis and quality review | None (reads image files directly) |
+| [**create**](create/SKILL.md) | Text-to-image generation | Any provider (see below) |
+| [**alter**](alter/SKILL.md) | Image modification (img2img, inpainting) | Any provider (see below) |
+
+### When To Use Each
+
+- **"What is in this image?"** — review
+- **"Generate a new image of..."** — create
+- **"Change this image to..."** — alter
+- **"Is this image ready for production?"** — review (quality gates)
+- **"Make the sky more dramatic"** — alter (img2img or inpainting)
+
+## The Creative Loop: Analyze-Execute-Review-Refine
+
+Every visual task follows this iterative cycle until the output matches requirements:
+
+```
+┌─────────┐     ┌─────────┐     ┌─────────┐     ┌─────────┐
+│ ANALYZE │────>│ EXECUTE │────>│ REVIEW  │────>│ REFINE  │
+│ (review)│     │(create/ │     │(review) │     │ (alter) │
+│         │     │ alter)  │     │         │     │         │
+└─────────┘     └─────────┘     └─────────┘     └────┬────┘
+     ^                                                │
+     └────────────────────────────────────────────────┘
+```
+
+1. **Analyze** — Understand the visual context, reference images, or requirements (review sub-skill)
+2. **Execute** — Generate new assets (create) or modify existing ones (alter)
+3. **Review** — Run quality gates: brand, accessibility, production, sensitivity (review sub-skill)
+4. **Refine** — Adjust prompts, parameters, or masks based on review findings (alter sub-skill)
+
+## Provider Status
+
+The imagery skill uses a provider abstraction layer. Review works without any provider. Create and alter require at least one:
+
+```bash
+# Check available providers
+sh "${CLAUDE_PLUGIN_ROOT}/scripts/_python.sh" "${CLAUDE_PLUGIN_ROOT}/skills/product/imagery/scripts/provider.py" detect
+```
+
+| Provider | How to Enable |
+|----------|---------------|
+| **cstudio** | Install CLI binary, set `GOOGLE_CLOUD_PROJECT` |
+| **vertex-curl** | `gcloud auth login`, set `GOOGLE_CLOUD_PROJECT` |
+| **openai** | Set `OPENAI_API_KEY` |
+| **stability** | Set `STABILITY_API_KEY` |
+| **replicate** | Set `REPLICATE_API_TOKEN` |
+
+## Quick Start
+
+### Analyze an existing image
+```
+Read the image file → apply review sub-skill analysis lenses
+```
+
+### Generate a new image
+```
+Describe requirements → create sub-skill generates → review sub-skill validates
+```
+
+### Modify an existing image
+```
+Review identifies issues → alter sub-skill applies changes → review validates
+```
+
+## Reference Map
+
+### Review References
+- `review/refs/analysis-lenses.md` — Four analysis lenses (general, technical, style, layout)
+- `review/refs/quality-gates.md` — Four quality gates (brand, a11y, production, sensitivity)
+
+### Create References
+- `create/refs/provider-reference.md` — Provider APIs and configuration
+- `create/refs/prompt-engineering.md` — Crafting effective prompts
+
+### Alter References
+- `alter/refs/editing-reference.md` — Modification patterns, provider commands, refinement strategies

package/skills/product/imagery/alter/SKILL.md

@@ -0,0 +1,105 @@
+---
+name: alter
+description: |
+  AI-powered image modification: img2img editing and mask-based inpainting.
+  Requires a provider that supports editing operations.
+
+  Use when: "edit image", "modify image", "change image", "inpaint", "img2img"
+phase_relevance: ["clarify", "design", "review"]
+archetype_relevance: ["*"]
+---
+
+# Image Alteration
+
+AI-powered modification of existing images through two primary modes: image-to-image editing (global changes) and mask-based inpainting (local changes).
+
+## When To Use This Skill
+
+- Modifying style, lighting, or mood of an existing image
+- Replacing or adding specific elements within an image
+- Iterating on generated images to reach desired output
+- Making targeted local edits while preserving the rest
+
+## Modification Modes
+
+### Image-to-Image (Global Edits)
+
+Changes the overall style or adds global elements while preserving core composition.
+
+```bash
+sh "${CLAUDE_PLUGIN_ROOT}/scripts/_python.sh" "${CLAUDE_PLUGIN_ROOT}/skills/imagery/scripts/provider.py" edit \
+  --image ./source.png \
+  --prompt "Same scene but with a dramatic sunset sky" \
+  --output ./v2.png
+```
+
+**Key control**: The strength parameter determines how much the model deviates from the original.
+- `--strength 0.2-0.3` — Subtle changes (color grading, minor adjustments)
+- `--strength 0.4-0.6` — Moderate changes (style transfer, lighting shifts)
+- `--strength 0.7-0.9` — Dramatic changes (major style overhaul)
+
+### Inpainting (Local Edits)
+
+Precise edits to specific regions using a binary mask (white = edit area, black = preserve).
+
+```bash
+sh "${CLAUDE_PLUGIN_ROOT}/scripts/_python.sh" "${CLAUDE_PLUGIN_ROOT}/skills/imagery/scripts/provider.py" inpaint \
+  --image ./source.png \
+  --mask ./mask.png \
+  --prompt "Replace with floating lanterns" \
+  --output ./v2_inpaint.png
+```
+
+**Best practices for masks:**
+- Clean edges — avoid feathered or anti-aliased mask boundaries
+- Include enough context area around the edit target
+- For complex shapes, create the mask programmatically or use a drawing tool
+
+## Iterative Refinement Loop
+
+When the output is close but not quite right:
+
+```
+1. DIAGNOSE  — What specifically is wrong? (color, composition, detail, artifact)
+2. ADJUST    — Modify the right parameter for the issue:
+   - Wrong elements → refine prompt, add negative prompt
+   - Wrong style → adjust strength parameter
+   - Wrong detail → fix seed, tweak guidance scale
+   - Local issue → switch to inpainting with a targeted mask
+3. REGENERATE — Run with adjusted parameters
+4. REVIEW     — Use the review sub-skill to validate
+```
+
+### Refinement Strategies
+
+| Problem | Solution |
+|---------|----------|
+| Missing element | Increase weight in prompt with specific adjectives |
+| Unwanted artifact | Add to `--negative-prompt` |
+| Good composition, bad detail | Fix `--seed`, adjust `--guidance-scale` |
+| Almost perfect, one area wrong | Switch to inpainting for that region |
+| Too different from original | Lower `--strength` value |
+| Not different enough | Raise `--strength` value |
+
+### Final Polish
+
+Use upscaling as the last step to bring a draft to production quality:
+
+```bash
+# Upscaling (cstudio only — other providers may not support this)
+sh "${CLAUDE_PLUGIN_ROOT}/scripts/_python.sh" "${CLAUDE_PLUGIN_ROOT}/skills/imagery/scripts/provider.py" generate \
+  --provider cstudio \
+  --prompt "upscale" \
+  --output ./final_hires.png
+```
+
+## Detailed References
+
+- **Editing techniques, provider commands, and refinement strategies**: [refs/editing-reference.md](refs/editing-reference.md)
+
+## Integration With Other Sub-Skills
+
+- Use **review** before altering to understand what needs to change
+- Use **review** after altering to validate the changes
+- If starting from scratch is better than editing, hand off to **create**
+- Alteration is the "Execute" and "Refine" steps of the Analyze-Execute-Review-Refine loop

package/skills/product/imagery/alter/refs/editing-reference.md

@@ -0,0 +1,146 @@
+---
+phase_relevance: ["clarify", "design", "review"]
+archetype_relevance: ["*"]
+---
+# Image Editing Reference
+
+Comprehensive reference for image modification techniques, provider-specific commands, and refinement strategies.
+
+---
+
+## Modification Patterns
+
+### Image-to-Image (Img2Img)
+
+**Use case:** Changing the overall style or adding global elements to an existing image while preserving the core composition.
+
+**Strength Parameter:** Controls how much the model deviates from the original:
+| Strength | Effect | When to Use |
+|----------|--------|-------------|
+| 0.1–0.3 | Subtle — color grading, minor adjustments | Fine-tuning a nearly-finished asset |
+| 0.4–0.6 | Moderate — style transfer, lighting shifts | Changing mood or atmosphere |
+| 0.7–0.9 | Dramatic — major style overhaul | Reimagining the visual entirely |
+
+**Prompting tip:** Focus on the *changes* you want to see, but reiterate the core elements you want to *keep*.
+
+### Inpainting & Masking
+
+**Use case:** Precise local edits (e.g., "Change the color of the car," "Add a person to the bench").
+
+**Mask creation:**
+- Binary mask: white = edit area, black = preserve
+- Clean edges — avoid feathered or anti-aliased mask boundaries
+- Include enough context area around the edit target (10-20% padding)
+- For complex shapes, create the mask programmatically or use a drawing tool
+
+**In-place consistency:** Inpainting maintains background and subject consistency while swapping specific details. Works best when the replacement content is stylistically compatible with the surrounding area.
+
+---
+
+## Refinement Strategies
+
+When the output is "almost there" but not quite:
+
+| Problem | Strategy | Implementation |
+|---------|----------|----------------|
+| Missing element | Increase weight in prompt | Add specific adjectives: "vibrant emerald green" not just "green" |
+| Unwanted artifact | Add negative constraints | `--negative-prompt "text, watermarks, blur, distortion"` |
+| Good composition, bad detail | Fix seed, tweak guidance | Lock `--seed`, adjust `--guidance-scale` up/down |
+| Almost perfect, one area wrong | Switch to inpainting | Create a mask for just that region |
+| Too different from original | Lower strength | Decrease `--strength` by 0.1-0.2 |
+| Not different enough | Raise strength | Increase `--strength` by 0.1-0.2 |
+| Wrong color balance | Adjust prompt specifics | Be explicit about lighting and color temperature |
+| Artifacts at edges | Better mask | Add padding around edit area, use cleaner mask edges |
+
+### Iteration Flow
+```
+1. DIAGNOSE  — What specifically is wrong? (color, composition, detail, artifact)
+2. ADJUST    — Modify the right parameter for the issue type
+3. REGENERATE — Run with adjusted parameters
+4. REVIEW     — Use the review sub-skill to validate
+5. REPEAT     — Until quality gates pass
+```
+
+---
+
+## Provider-Specific Commands
+
+### Provider-Agnostic (Recommended)
+
+```bash
+# Image-to-image edit
+sh "${CLAUDE_PLUGIN_ROOT}/scripts/_python.sh" "${CLAUDE_PLUGIN_ROOT}/skills/imagery/scripts/provider.py" edit \
+  --image ./source.png \
+  --prompt "Same scene but with dramatic sunset sky" \
+  --output ./v2.png
+
+# Inpainting with mask
+sh "${CLAUDE_PLUGIN_ROOT}/scripts/_python.sh" "${CLAUDE_PLUGIN_ROOT}/skills/imagery/scripts/provider.py" inpaint \
+  --image ./source.png \
+  --mask ./mask.png \
+  --prompt "Replace with floating lanterns" \
+  --output ./v2_inpaint.png
+
+# Use a specific provider
+sh "${CLAUDE_PLUGIN_ROOT}/scripts/_python.sh" "${CLAUDE_PLUGIN_ROOT}/skills/imagery/scripts/provider.py" edit \
+  --provider stability \
+  --image ./source.png \
+  --prompt "Oil painting style" \
+  --output ./v2_oil.png
+```
+
+### cstudio CLI (Vertex AI Creative Studio)
+
+```bash
+# Configuration
+export GOOGLE_CLOUD_PROJECT="your-project-id"
+export GOOGLE_CLOUD_LOCATION="us-central1"
+
+# Image-to-image
+cstudio edit image \
+  --image ./v1.png \
+  --prompt "A gothic garden with purple roses" \
+  --output ./v2.png
+
+# Inpainting
+cstudio inpaint image \
+  --image ./v1.png \
+  --mask ./mask.png \
+  --prompt "Replace statues with floating lanterns" \
+  --output ./v1_inpaint.png
+
+# Upscaling (cstudio only)
+cstudio upscale image --image ./draft.png --output ./final_hires.png
+
+# Advanced controls
+cstudio edit image \
+  --image ./v1.png \
+  --prompt "Same scene, golden hour lighting" \
+  --negative-prompt "text, watermarks" \
+  --seed 42 \
+  --guidance-scale 8.5 \
+  --output ./v2_golden.png
+```
+
+### Advanced Controls (All Providers)
+
+| Parameter | Purpose | Example |
+|-----------|---------|---------|
+| `--prompt` | What to change/add | "Dramatic sunset sky" |
+| `--negative-prompt` | Elements to exclude | "text, watermarks, blur" |
+| `--seed` | Reproducible results | Any integer |
+| `--guidance-scale` | Prompt adherence | Higher = more literal |
+| `--strength` | Deviation from original (img2img) | 0.1–0.9 |
+| `--sample-count` | Multiple variations | 1–4 (cstudio only) |
+
+---
+
+## Final Polish
+
+Use upscaling as the last step to bring a draft to production quality. Currently only available via cstudio:
+
+```bash
+cstudio upscale image --image ./final_draft.png --output ./final_hires.png
+```
+
+For other providers, generate at the highest resolution supported and use external tools for additional upscaling if needed.

package/skills/product/imagery/create/SKILL.md

@@ -0,0 +1,109 @@
+---
+name: create
+description: |
+  AI-powered image generation from text prompts using multiple providers.
+  Supports 5 providers: cstudio, vertex-curl, OpenAI, Stability AI, Replicate.
+
+  Use when: "generate image", "create image", "text to image", "new visual"
+phase_relevance: ["clarify", "design", "review"]
+archetype_relevance: ["*"]
+---
+
+# Image Creation
+
+AI-powered image generation from text prompts. Supports multiple providers through a unified abstraction layer.
+
+## When To Use This Skill
+
+- Generating new images from text descriptions
+- Creating visual assets for marketing, UI, presentations
+- Producing multiple variations of a concept
+- Building visual prototypes from written specifications
+
+## Supported Providers
+
+| Provider | Interface | Auth | Best For |
+|----------|-----------|------|----------|
+| **cstudio** | CLI binary | `GOOGLE_CLOUD_PROJECT` | Interactive use, rapid iteration |
+| **vertex-curl** | gcloud + curl | `gcloud auth` | CI/CD, scripted pipelines |
+| **openai** | REST API | `OPENAI_API_KEY` | Existing OpenAI subscription |
+| **stability** | REST API | `STABILITY_API_KEY` | Stable Diffusion 3.5 |
+| **replicate** | REST API | `REPLICATE_API_TOKEN` | Flux models, pay-per-use |
+
+### Provider Detection
+
+```bash
+# Check which providers are available
+sh "${CLAUDE_PLUGIN_ROOT}/scripts/_python.sh" "${CLAUDE_PLUGIN_ROOT}/skills/imagery/scripts/provider.py" detect
+```
+
+The provider abstraction selects the best available provider automatically. A specific provider can be forced when needed.
+
+## Generation Workflow
+
+```
+1. UNDERSTAND — Clarify the visual intent and requirements
+2. CRAFT     — Build a high-quality prompt with style and constraints
+3. GENERATE  — Execute via the provider abstraction
+4. REVIEW    — Use the review sub-skill to validate output
+5. REFINE    — Iterate on prompt, parameters, or provider settings
+```
+
+### Step 1: Understand Requirements
+
+Before generating, gather:
+- **Subject**: What is the image of?
+- **Style**: Photorealistic, illustration, abstract, etc.
+- **Dimensions**: Aspect ratio (16:9, 1:1, 9:16, etc.)
+- **Context**: Where will this image be used?
+- **Constraints**: Brand colors, no text, specific mood
+
+### Step 2: Craft the Prompt
+
+Effective prompts follow this structure:
+
+```
+[Subject] + [Style/Medium] + [Lighting/Mood] + [Composition] + [Details]
+```
+
+Example: "A minimalist office workspace with a single monitor, soft natural side lighting, centered composition, matte finish, neutral color palette"
+
+For detailed prompt engineering techniques, see [refs/prompt-engineering.md](refs/prompt-engineering.md).
+
+### Step 3: Generate
+
+```bash
+# Basic generation (auto-detects best available provider)
+sh "${CLAUDE_PLUGIN_ROOT}/scripts/_python.sh" "${CLAUDE_PLUGIN_ROOT}/skills/imagery/scripts/provider.py" generate \
+  --prompt "Your crafted prompt" \
+  --output ./output/v1.png
+
+# Use a specific provider
+sh "${CLAUDE_PLUGIN_ROOT}/scripts/_python.sh" "${CLAUDE_PLUGIN_ROOT}/skills/imagery/scripts/provider.py" generate \
+  --prompt "Your crafted prompt" \
+  --provider openai \
+  --output ./output/v1.png
+```
+
+### Key Parameters
+
+| Parameter | Purpose | Example |
+|-----------|---------|---------|
+| `--prompt` | Text description | "A gothic garden at dusk" |
+| `--aspect-ratio` | Output dimensions | "16:9", "1:1", "9:16" |
+| `--negative-prompt` | Elements to exclude | "text, watermarks, blur" |
+| `--sample-count` | Number of variations | 1-4 |
+| `--seed` | Reproducible results | Any integer |
+| `--guidance-scale` | Prompt adherence | Higher = more literal |
+| `--model` | Model selection | "imagen-3.0-generate-001" |
+
+## Detailed References
+
+- **Provider APIs and configuration**: [refs/provider-reference.md](refs/provider-reference.md)
+- **Prompt engineering techniques**: [refs/prompt-engineering.md](refs/prompt-engineering.md)
+
+## Integration With Other Sub-Skills
+
+- Use **review** after generation to validate quality and run gates
+- If modifications are needed, hand off to **alter** for targeted edits
+- Generation output feeds into the Analyze-Execute-Review-Refine loop