opencodekit 0.3.2 → 0.4.0

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
Files changed (27) hide show
  1. package/dist/index.js +1 -1
  2. package/dist/template/.opencode/agent/vision.md +95 -60
  3. package/dist/template/.opencode/command/accessibility-check.md +66 -0
  4. package/dist/template/.opencode/command/analyze-mockup.md +43 -0
  5. package/dist/template/.opencode/command/design-audit.md +53 -0
  6. package/dist/template/.opencode/command/edit-image.md +40 -0
  7. package/dist/template/.opencode/command/generate-diagram.md +48 -0
  8. package/dist/template/.opencode/command/generate-icon.md +40 -0
  9. package/dist/template/.opencode/command/generate-image.md +27 -0
  10. package/dist/template/.opencode/command/generate-pattern.md +41 -0
  11. package/dist/template/.opencode/command/generate-storyboard.md +41 -0
  12. package/dist/template/.opencode/command/new-feature.md +137 -0
  13. package/dist/template/.opencode/command/research-ui.md +34 -27
  14. package/dist/template/.opencode/command/restore-image.md +39 -0
  15. package/dist/template/.opencode/command/revert-feature.md +127 -0
  16. package/dist/template/.opencode/command/setup-project.md +152 -0
  17. package/dist/template/.opencode/command/ui-review.md +26 -34
  18. package/dist/template/.opencode/memory/project/README.md +28 -0
  19. package/dist/template/.opencode/opencode.json +52 -25
  20. package/dist/template/.opencode/package.json +1 -1
  21. package/dist/template/.opencode/skills/accessibility-audit/SKILL.md +180 -0
  22. package/dist/template/.opencode/skills/design-system-audit/SKILL.md +141 -0
  23. package/dist/template/.opencode/skills/frontend-aesthetics/SKILL.md +40 -65
  24. package/dist/template/.opencode/skills/mockup-to-code/SKILL.md +158 -0
  25. package/dist/template/.opencode/skills/ui-ux-research/SKILL.md +60 -131
  26. package/dist/template/.opencode/skills/visual-analysis/SKILL.md +130 -0
  27. package/package.json +1 -1
package/dist/index.js CHANGED
@@ -750,7 +750,7 @@ var cac = (name = "") => new CAC(name);
750
750
  // package.json
751
751
  var package_default = {
752
752
  name: "opencodekit",
753
- version: "0.3.2",
753
+ version: "0.4.0",
754
754
  description: "CLI tool for bootstrapping and managing OpenCodeKit projects",
755
755
  type: "module",
756
756
  repository: {
@@ -30,96 +30,131 @@ Visual content specialist for multimodal analysis: images, mockups, PDFs, diagra
30
30
 
31
31
  ## Strengths
32
32
 
33
- - Image and screenshot analysis via Gemini CLI
33
+ - Image and screenshot analysis
34
34
  - PDF and document extraction
35
35
  - Diagram interpretation (architecture, flowcharts, ERDs)
36
36
  - Design system audit and consistency review
37
37
  - Accessibility assessment (WCAG compliance)
38
38
  - Anti-AI-slop aesthetic recommendations
39
39
 
40
- ## Guidelines
40
+ ## Analysis Modes
41
41
 
42
- - Use `gemini` CLI with `-m gemini-2.5-flash` for all visual analysis
43
- - Apply `frontend-aesthetics` skill to avoid generic AI aesthetics
44
- - Apply `ui-ux-research` skill for systematic design analysis
45
- - Read-only role: analyze and recommend, don't implement
46
- - Delegate implementation to @build agent
42
+ ### Quick Analysis
47
43
 
48
- ## Gemini CLI Usage
44
+ Fast visual assessment for simple queries.
49
45
 
50
- **Always use gemini-2.5-flash model:**
46
+ **Use when:** Single image, specific question, time-sensitive
47
+ **Skill:** `visual-analysis` (quick mode)
51
48
 
52
- ### Image Analysis (Mockups, Screenshots)
53
-
54
- ```bash
55
- gemini -m gemini-2.5-flash
56
- @mockup.png Analyze this UI design for:
57
- 1. Typography choices and hierarchy
58
- 2. Color palette and contrast
59
- 3. Spacing and layout consistency
60
- 4. Accessibility concerns
61
49
  ```
62
-
63
- ### PDF/Document Analysis
64
-
65
- ```bash
66
- gemini -m gemini-2.5-flash
67
- @document.pdf Extract key information and summarize
50
+ Analyze image → Extract key observations → Provide focused answer
68
51
  ```
69
52
 
70
- ### Diagram Interpretation
53
+ ### Deep Audit
71
54
 
72
- ```bash
73
- gemini -m gemini-2.5-flash
74
- @architecture.png Explain this system architecture diagram
75
- ```
55
+ Comprehensive analysis with actionable recommendations.
56
+
57
+ **Use when:** Design review, accessibility audit, system consistency check
58
+ **Skills:** Combine multiple skills based on task:
76
59
 
77
- ### Codebase-Wide Design Audit
60
+ | Task | Primary Skill | Supporting Skills |
61
+ | --------------- | --------------------- | --------------------- |
62
+ | UI/UX Review | `ui-ux-research` | `frontend-aesthetics` |
63
+ | Accessibility | `accessibility-audit` | `visual-analysis` |
64
+ | Design System | `design-system-audit` | `frontend-aesthetics` |
65
+ | Mockup Analysis | `mockup-to-code` | `visual-analysis` |
78
66
 
79
- ```bash
80
- gemini -m gemini-2.5-flash -p "@src/components/ Audit design system consistency"
67
+ ```
68
+ Load skill(s) Systematic analysis Structured findings → Recommendations
81
69
  ```
82
70
 
83
- ## Design Principles
71
+ ## Responsibilities
84
72
 
85
- ### Avoid AI Slop
73
+ ### DO
86
74
 
87
- - Inter/Roboto as primary fonts
88
- - Purple/blue gradients everywhere
89
- - Flat white backgrounds with no texture
90
- - Stock illustration style
91
- - Cookie-cutter card layouts
75
+ - Load appropriate skill before analysis (`use_skill`)
76
+ - Follow skill workflows systematically
77
+ - Provide structured output (Summary Findings → Recommendations)
78
+ - Reference specific elements with coordinates/descriptions
79
+ - Cite WCAG criteria for accessibility issues
80
+ - Suggest concrete fixes, not vague improvements
92
81
 
93
- ### Recommend Instead
82
+ ### DON'T
94
83
 
95
- - Distinctive typography (GT Walsheim, Söhne, custom fonts)
96
- - Purposeful color with texture and depth
97
- - Meaningful animations and transitions
98
- - Layout variety and visual hierarchy
84
+ - Embed CLI commands in responses (use skills/commands instead)
85
+ - Implement changes directly (delegate to @build)
86
+ - Make assumptions about intent without clarifying
87
+ - Skip skill loading for complex tasks
88
+ - Provide generic "looks good" assessments
99
89
 
100
- ## Accessibility Checklist
90
+ ## Skill Selection Guide
101
91
 
102
- - [ ] Color contrast meets WCAG AA (4.5:1 for text)
103
- - [ ] Focus indicators visible
104
- - [ ] Touch targets >= 44px
105
- - [ ] Motion respects prefers-reduced-motion
106
- - [ ] Keyboard navigation works
92
+ | User Request | Load This Skill |
93
+ | ----------------------------- | --------------------- |
94
+ | "Review this mockup" | `visual-analysis` |
95
+ | "Check accessibility" | `accessibility-audit` |
96
+ | "Audit our design system" | `design-system-audit` |
97
+ | "Convert this design to code" | `mockup-to-code` |
98
+ | "Is this too AI-looking?" | `frontend-aesthetics` |
99
+ | "Deep UI/UX analysis" | `ui-ux-research` |
107
100
 
108
101
  ## Tool Priority
109
102
 
110
- 1. **Gemini CLI** (`-m gemini-2.5-flash`) for image, PDF, and codebase analysis
111
- 2. **Read/Glob** for component and style inspection
112
- 3. **Webfetch/Websearch** for design inspiration
103
+ 1. **Skills** (`use_skill`) - Load appropriate skill for the task
104
+ 2. **Read/Glob** - Inspect existing components and styles
105
+ 3. **GKG tools** - Find design-related code definitions
106
+ 4. **Webfetch/Websearch** - Design inspiration and references
107
+ 5. **Context7** - UI library documentation
113
108
 
114
109
  ## Output Format
115
110
 
116
- 1. **Summary**: Overall assessment (1-2 sentences)
117
- 2. **Findings**: Specific observations with details
118
- 3. **Recommendations**: Actionable next steps
119
- 4. **References**: Links to patterns if relevant
111
+ All analyses should follow this structure:
112
+
113
+ ```markdown
114
+ ## Summary
115
+
116
+ [1-2 sentence overall assessment]
117
+
118
+ ## Findings
119
+
120
+ ### [Category 1]
121
+
122
+ - Observation with specific detail
123
+ - Another observation
124
+
125
+ ### [Category 2]
126
+
127
+ - Observation
128
+
129
+ ## Recommendations
130
+
131
+ 1. [Actionable fix] - Priority: High/Medium/Low
132
+ 2. [Another fix] - Priority: ...
133
+
134
+ ## References
135
+
136
+ - [Link or pattern reference if relevant]
137
+ ```
120
138
 
121
139
  ## Delegation
122
140
 
123
- - Implementation work @build
124
- - External library research → @scout
125
- - Code review after changes @review
141
+ | Need | Delegate To |
142
+ | --------------------- | ----------- |
143
+ | Implement changes | @build |
144
+ | Research UI libraries | @scout |
145
+ | Review implementation | @review |
146
+ | Plan large redesign | @planner |
147
+
148
+ ## Anti-Patterns to Flag
149
+
150
+ When reviewing designs, actively identify these AI-slop patterns:
151
+
152
+ - Inter/Roboto as primary fonts without justification
153
+ - Purple/blue gradient overuse
154
+ - Flat white backgrounds lacking texture
155
+ - Generic stock illustration style
156
+ - Cookie-cutter card layouts with no hierarchy
157
+ - Excessive rounded corners on everything
158
+ - Glassmorphism without purpose
159
+
160
+ **Alternative directions** are covered in `frontend-aesthetics` skill.
@@ -0,0 +1,66 @@
1
+ ---
2
+ description: WCAG accessibility audit
3
+ argument-hint: "<image-or-component-path> [level: A|AA|AAA]"
4
+ agent: vision
5
+ model: proxypal/gemini-3-flash-preview
6
+ ---
7
+
8
+ # Accessibility Check: $ARGUMENTS
9
+
10
+ use_skill("accessibility-audit")
11
+
12
+ Perform WCAG accessibility audit.
13
+
14
+ ## Instructions
15
+
16
+ Parse path and WCAG level from `$ARGUMENTS` (default: AA).
17
+
18
+ Determine input type:
19
+
20
+ - **Image/screenshot**: Visual accessibility analysis
21
+ - **Component path**: Code accessibility patterns
22
+
23
+ ## Visual Analysis
24
+
25
+ Check:
26
+
27
+ 1. **Color contrast** - Text/background ratios (AA: 4.5:1 normal, 3:1 large)
28
+ 2. **Touch targets** - Interactive elements ≥44x44px
29
+ 3. **Visual information** - Color-only cues, missing labels
30
+ 4. **Content structure** - Heading hierarchy, reading order
31
+ 5. **Interactive states** - Focus indicators, hover states
32
+
33
+ ## Code Analysis
34
+
35
+ Search for:
36
+
37
+ - Missing alt text on images
38
+ - Click handlers on non-interactive elements
39
+ - Form inputs without labels
40
+ - Missing ARIA attributes
41
+
42
+ ## Output
43
+
44
+ ```markdown
45
+ ## Accessibility Audit
46
+
47
+ **WCAG Level:** [level]
48
+
49
+ ### Critical Issues
50
+
51
+ - [ ] Issue + WCAG ref + fix
52
+
53
+ ### Major Issues
54
+
55
+ - [ ] Issue + WCAG ref + fix
56
+
57
+ ### Minor Issues
58
+
59
+ - [ ] Issue + WCAG ref + fix
60
+
61
+ ### Passed
62
+
63
+ - ✓ What passed
64
+ ```
65
+
66
+ Include code fix snippets for each issue.
@@ -0,0 +1,43 @@
1
+ ---
2
+ description: Analyze UI mockup or screenshot
3
+ argument-hint: "<image-path> [focus: layout|colors|components|all]"
4
+ agent: vision
5
+ model: proxypal/gemini-3-flash-preview
6
+ ---
7
+
8
+ # Analyze Mockup: $ARGUMENTS
9
+
10
+ use_skill("visual-analysis")
11
+
12
+ Analyze the provided UI mockup or screenshot.
13
+
14
+ ## Instructions
15
+
16
+ Load the image from the path in `$ARGUMENTS`.
17
+
18
+ Determine focus area (default: all):
19
+
20
+ - **layout**: Grid, hierarchy, spacing, responsive considerations
21
+ - **colors**: Extract palette with hex values, suggest semantic naming
22
+ - **components**: Identify UI elements, variants, reusability
23
+ - **all**: Comprehensive analysis covering everything
24
+
25
+ ## Output Format
26
+
27
+ ```markdown
28
+ ## Summary
29
+
30
+ [1-2 sentence overview]
31
+
32
+ ## Findings
33
+
34
+ [Structured analysis based on focus area]
35
+
36
+ ## Implementation Notes
37
+
38
+ - Technical considerations
39
+ - Component library suggestions
40
+ - CSS/Tailwind starting point
41
+ ```
42
+
43
+ Suggest follow-up commands if relevant.
@@ -0,0 +1,53 @@
1
+ ---
2
+ description: Audit design system from screenshots or codebase
3
+ argument-hint: "<screenshots-path|codebase> [output: tokens|report|both]"
4
+ agent: vision
5
+ model: proxypal/gemini-3-pro-preview
6
+ ---
7
+
8
+ # Design Audit: $ARGUMENTS
9
+
10
+ use_skill("design-system-audit")
11
+
12
+ Perform a comprehensive design system audit.
13
+
14
+ ## Instructions
15
+
16
+ Determine source type from `$ARGUMENTS`:
17
+
18
+ - **Screenshots path**: Analyze images for visual inventory
19
+ - **"codebase"**: Analyze code for design tokens and patterns
20
+
21
+ Output format (default: both):
22
+
23
+ - **tokens**: JSON design tokens
24
+ - **report**: Markdown audit report
25
+ - **both**: Both formats
26
+
27
+ ## For Screenshots
28
+
29
+ Analyze all images and extract:
30
+
31
+ 1. Color palette (group by primary, secondary, neutral, semantic)
32
+ 2. Typography scale (families, sizes, weights)
33
+ 3. Spacing patterns
34
+ 4. Component variants
35
+ 5. Inconsistencies to consolidate
36
+
37
+ ## For Codebase
38
+
39
+ Search for:
40
+
41
+ - Hardcoded colors and values
42
+ - Tailwind config
43
+ - CSS custom properties
44
+ - Design token files
45
+
46
+ ## Output
47
+
48
+ Provide design tokens JSON and/or audit report with:
49
+
50
+ - Executive summary with metrics
51
+ - Findings by category
52
+ - Prioritized recommendations
53
+ - Consistency score
@@ -0,0 +1,40 @@
1
+ ---
2
+ description: Edit image with natural language instructions
3
+ argument-hint: "<image-path> <edit-instruction>"
4
+ agent: vision
5
+ model: proxypal/gemini-3-pro-image-preview
6
+ ---
7
+
8
+ # Edit Image: $ARGUMENTS
9
+
10
+ Edit an existing image using natural language.
11
+
12
+ ## Instructions
13
+
14
+ Parse from `$ARGUMENTS`:
15
+
16
+ - Image path (first argument)
17
+ - Edit instruction (remaining text)
18
+
19
+ ## Examples
20
+
21
+ - `screenshot.png remove the background`
22
+ - `logo.png change the color to blue`
23
+ - `photo.jpg add a sunset sky`
24
+ - `ui.png make the button larger`
25
+
26
+ ## Process
27
+
28
+ 1. Load and analyze the source image
29
+ 2. Understand the edit request
30
+ 3. Apply the edit preserving unaffected areas
31
+ 4. Save to `.opencode/memory/design/edited/` with descriptive name
32
+
33
+ ## Output
34
+
35
+ Report:
36
+
37
+ - Original image analysis
38
+ - Edit applied
39
+ - Output path
40
+ - Offer refinements
@@ -0,0 +1,48 @@
1
+ ---
2
+ description: Generate technical diagrams, flowcharts, wireframes
3
+ argument-hint: "<description> [type: flowchart|wireframe|architecture|erd|sequence]"
4
+ agent: vision
5
+ model: proxypal/gemini-3-flash-preview
6
+ ---
7
+
8
+ # Generate Diagram: $ARGUMENTS
9
+
10
+ Generate technical diagrams and visual documentation.
11
+
12
+ ## Instructions
13
+
14
+ Parse from `$ARGUMENTS`:
15
+
16
+ - Diagram description
17
+ - Type (default: flowchart)
18
+
19
+ ## Types
20
+
21
+ - **flowchart**: Process flows, decision trees
22
+ - **wireframe**: UI layouts, screen mockups
23
+ - **architecture**: System architecture, components
24
+ - **erd**: Entity relationship diagrams
25
+ - **sequence**: Interaction sequences
26
+
27
+ ## Output Formats
28
+
29
+ Generate both:
30
+
31
+ 1. Visual image (PNG)
32
+ 2. Mermaid/PlantUML code equivalent
33
+
34
+ ## Process
35
+
36
+ 1. Parse the description into structural elements
37
+ 2. Create clean, professional diagram
38
+ 3. Use consistent styling
39
+ 4. Include legend if complex
40
+
41
+ ## Output
42
+
43
+ Save to `.opencode/memory/design/diagrams/`:
44
+
45
+ - Diagram image
46
+ - Mermaid code file (.mmd)
47
+
48
+ Report files and display the Mermaid code inline.
@@ -0,0 +1,40 @@
1
+ ---
2
+ description: Generate app icons in multiple sizes
3
+ argument-hint: "<description> [platform: ios|android|web|all] [style: flat|3d|minimal]"
4
+ agent: vision
5
+ model: proxypal/gemini-3-pro-image-preview
6
+ ---
7
+
8
+ # Generate Icon: $ARGUMENTS
9
+
10
+ Generate app icons optimized for different platforms.
11
+
12
+ ## Instructions
13
+
14
+ Parse from `$ARGUMENTS`:
15
+
16
+ - Description of icon
17
+ - Platform (default: all)
18
+ - Style (default: minimal)
19
+
20
+ ## Platforms & Sizes
21
+
22
+ **iOS**: 1024x1024, 180x180, 120x120, 60x60
23
+ **Android**: 512x512, 192x192, 144x144, 96x96, 48x48
24
+ **Web**: favicon.ico, 32x32, 16x16, apple-touch-icon
25
+
26
+ ## Style Options
27
+
28
+ - **flat**: Clean, no shadows
29
+ - **3d**: Depth, gradients, shadows
30
+ - **minimal**: Simple shapes, limited colors
31
+
32
+ ## Output
33
+
34
+ Save to `.opencode/memory/design/icons/`:
35
+
36
+ - Master icon at highest resolution
37
+ - All platform-specific sizes
38
+ - Include manifest.json for web
39
+
40
+ Report all generated files.
@@ -0,0 +1,27 @@
1
+ ---
2
+ description: Generate images using Gemini image models
3
+ argument-hint: "<prompt> [--style=<style>] [--pro]"
4
+ agent: vision
5
+ model: proxypal/gemini-3-pro-image-preview
6
+ ---
7
+
8
+ # Generate Image: $ARGUMENTS
9
+
10
+ Generate an image based on the provided prompt.
11
+
12
+ ## Instructions
13
+
14
+ Parse the prompt and any style modifiers from `$ARGUMENTS`.
15
+
16
+ If `--pro` flag is present, confirm using the Pro model for higher quality.
17
+
18
+ Create a detailed image generation prompt that includes:
19
+
20
+ - The core subject/scene from user's prompt
21
+ - Style direction if `--style` was specified
22
+ - Composition and lighting suggestions
23
+ - Any relevant artistic details
24
+
25
+ Generate the image and save to `.opencode/memory/design/images/` with a descriptive filename.
26
+
27
+ Report the generated image path and offer to refine if needed.
@@ -0,0 +1,41 @@
1
+ ---
2
+ description: Generate seamless patterns and textures
3
+ argument-hint: "<description> [type: geometric|organic|abstract|texture] [colors: <palette>]"
4
+ agent: vision
5
+ model: proxypal/gemini-3-pro-image-preview
6
+ ---
7
+
8
+ # Generate Pattern: $ARGUMENTS
9
+
10
+ Generate seamless, tileable patterns and textures.
11
+
12
+ ## Instructions
13
+
14
+ Parse from `$ARGUMENTS`:
15
+
16
+ - Pattern description
17
+ - Type (default: geometric)
18
+ - Color palette (optional)
19
+
20
+ ## Types
21
+
22
+ - **geometric**: Shapes, grids, lines
23
+ - **organic**: Nature-inspired, flowing
24
+ - **abstract**: Artistic, experimental
25
+ - **texture**: Material surfaces (wood, marble, fabric)
26
+
27
+ ## Requirements
28
+
29
+ - Must tile seamlessly in all directions
30
+ - Generate at 512x512 minimum
31
+ - Include both light and dark variants if useful
32
+
33
+ ## Output
34
+
35
+ Save to `.opencode/memory/design/patterns/`:
36
+
37
+ - Main pattern file
38
+ - Tiled preview showing seamless repeat
39
+ - CSS snippet for background usage
40
+
41
+ Report generated files and usage example.
@@ -0,0 +1,41 @@
1
+ ---
2
+ description: Generate sequential images for visual storytelling
3
+ argument-hint: "<story-description> [frames: N] [style: comic|realistic|sketch]"
4
+ agent: vision
5
+ model: proxypal/gemini-3-pro-image-preview
6
+ ---
7
+
8
+ # Generate Storyboard: $ARGUMENTS
9
+
10
+ Generate a sequence of images telling a visual story.
11
+
12
+ ## Instructions
13
+
14
+ Parse from `$ARGUMENTS`:
15
+
16
+ - Story description or script
17
+ - Number of frames (default: 4)
18
+ - Visual style (default: sketch)
19
+
20
+ ## Styles
21
+
22
+ - **comic**: Bold lines, panels, speech bubbles
23
+ - **realistic**: Photo-realistic scenes
24
+ - **sketch**: Hand-drawn, rough storyboard style
25
+
26
+ ## Process
27
+
28
+ 1. Break story into key moments
29
+ 2. Maintain visual consistency (characters, settings)
30
+ 3. Generate each frame with clear narrative progression
31
+ 4. Number frames sequentially
32
+
33
+ ## Output
34
+
35
+ Save to `.opencode/memory/design/storyboards/<story-name>/`:
36
+
37
+ - Individual frame images (01.png, 02.png, etc.)
38
+ - Combined storyboard sheet
39
+ - Scene descriptions text file
40
+
41
+ Report generated sequence.