@hienlh/ppm 0.9.81 → 0.9.82

package/.opencode/commands/ckm-write-publish.md

@@ -1,74 +0,0 @@
----
-description: "Audit content, auto-fix issues, output publish-ready version - Args: <file-path>"
----
-
-Quality gate before publishing. Audit → Fix → Output.
-
-<file_path>$ARGUMENTS</file_path>
-
-## Workflow
-
-1. **Read Content**
-   - Load file from `<file_path>`
-   - Detect type: social | article | email | infographic
-
-2. **Run Audit** (use `/write/audit` criteria)
-   - Score each dimension (Copywriting, SEO, Platform, Brand)
-   - Identify issues
-
-3. **Auto-Fix Issues**
-
-   | Issue Type | Fix Action |
-   |------------|------------|
-   | Weak hook | Rewrite using hook formulas from `social-media/references/hook-writing.md` |
-   | Missing hashtags | Add platform-appropriate hashtags |
-   | Too long | Trim to platform limits |
-   | Weak CTA | Apply CTA patterns from `copywriting/references/cta-patterns.md` |
-   | Low readability | Simplify to Grade 6 level |
-   | Missing @mentions | Suggest relevant accounts |
-
-4. **Output**
-   - Save fixed version → same path (overwrite) or `{path}-published.md`
-   - Show before/after diff for major changes
-   - Final score must be ≥8.0 to pass
-
-## Pass Criteria
-
-| Score | Status | Action |
-|-------|--------|--------|
-| ≥9.0 | Publish Ready | No changes needed |
-| 8.0-8.9 | Good | Minor fixes applied |
-| 6.0-7.9 | Needs Work | Major fixes applied |
-| <6.0 | Rewrite | Flag for manual review |
-
-## Skills Used
-
-- `copywriting` - Hook formulas, CTA patterns
-- `social-media` - Platform specs, hashtag strategy
-- `seo-optimization` - Keyword optimization
-- `brand-guidelines` - Voice consistency
-
-## Output Format
-
-```markdown
-# Publish Review: {filename}
-
-**Before:** X.X/10 → **After:** X.X/10
-
-## Changes Made
-1. ✅ [Platform] Hook strengthened: "..." → "..."
-2. ✅ [Platform] Added X hashtags
-3. ✅ [Platform] CTA improved
-
-## Final Version
-[Full polished content here]
-
----
-**Status:** Ready to Publish ✅
-```
-
-## Examples
-```
-/write/publish assets/posts/251225-rickroll-all-platforms.md
-/write/publish assets/articles/251225-seo-guide.md
-```

package/.opencode/commands/ckm-youtube-blog.md

@@ -1,136 +0,0 @@
----
-description: "Convert YouTube video to SEO-optimized blog post - Args: <youtube-url> [writing-style]"
----
-
-# YouTube to Blog Post
-
-Transform YouTube video into SEO-optimized long-form article.
-
-<youtube_url>$ARGUMENTS</youtube_url>
-
-## Workflow
-
-1. **Extract Video Content**
-   - Activate `youtube-handling` skill
-   - Run VidCap API:
-     ```bash
-     # Get full article conversion
-     python .opencode/skills/youtube-handling/scripts/vidcap.py article "<url>"
-
-     # Or custom extraction for more control
-     python .opencode/skills/youtube-handling/scripts/vidcap.py summary-custom "<url>" \
-       --prompt "Extract main points, key quotes, and actionable insights"
-
-     # Get captions for quotes
-     python .opencode/skills/youtube-handling/scripts/vidcap.py caption "<url>" --ext vtt
-     ```
-   - Capture: title, key points, timestamps, quotes, statistics
-
-2. **Apply Writing Style**
-   - Check for writing style in `assets/writing-styles/`
-   - If style specified, apply voice/tone characteristics
-   - Default: professional, educational tone
-   - Activate `copywriting` skill for conversion techniques
-
-3. **SEO Optimization**
-   - Activate `seo-optimization` skill
-   - Generate:
-     - Primary keyword from video topic
-     - Secondary keywords from content
-     - Meta description (150-160 chars)
-     - URL slug
-   - Structure headings (H1, H2, H3) for featured snippets
-
-4. **Content Enhancement**
-   - Add intro hook (PAS or AIDA formula)
-   - Convert timestamps to section headers
-   - Add internal linking opportunities
-   - Create conclusion with CTA
-   - Include video embed
-
-5. **Output**
-   - Blog post → `assets/articles/{date}-{slug}.md`
-   - Frontmatter includes:
-     - SEO metadata
-     - Source video URL
-     - Writing style used
-     - Keywords
-
-## Skills Used
-
-- `youtube-handling` - VidCap API for video extraction
-- `copywriting` - Writing formulas and style
-- `seo-optimization` - Keyword research, meta optimization
-- `content-marketing` - Content structure
-- `brand-guidelines` - Brand voice consistency
-- `assets-organizing` - Standardized output paths
-
-## Agents Used
-
-- `content-creator` - Article writing
-- `copywriter` - Hooks, CTAs
-- `seo-specialist` - Keyword optimization
-
-## Writing Styles
-
-Styles are defined in `assets/writing-styles/` or `references/writing-styles.md`:
-
-| Style | Characteristics |
-|-------|-----------------|
-| casual | Conversational, contractions, personal pronouns |
-| professional | Formal, authoritative, data-driven |
-| edgy | Provocative, bold claims, pattern interrupts |
-| luxe | Elegant, minimal, premium language |
-| warm | Supportive, empathetic, encouraging |
-
-## Output Format
-
-```yaml
----
-title: "SEO Optimized Title"
-slug: keyword-rich-slug
-description: "Meta description under 160 chars"
-keywords: [primary, secondary, related]
-source_video: "https://youtube.com/watch?v=..."
-writing_style: professional
-date: YYYY-MM-DD
-schema: Article
----
-
-# {Title}
-
-{Intro with hook}
-
-## {Section from video timestamp}
-
-{Content with quotes from video}
-
-> "Direct quote from video" - Speaker
-
-## Key Takeaways
-
-- Point 1
-- Point 2
-
-## Conclusion
-
-{CTA}
-
----
-*This article is based on [Video Title](source_url)*
-```
-
-## Quality Gate (Auto-Triggered)
-
-**After generating content, MUST run:**
-1. `/write/audit` → Score content (target: ≥8.0)
-2. `/write/publish` → Auto-fix if score <8.0
-3. Present final version with score
-
-## Examples
-
-```bash
-/youtube:blog "https://youtube.com/watch?v=xyz"
-/youtube:blog "https://youtu.be/abc" casual
-/youtube:blog "https://youtube.com/watch?v=123" professional
-```

package/.opencode/commands/ckm-youtube-infographic.md

@@ -1,156 +0,0 @@
----
-description: "Convert YouTube video to visual infographic - Args: <youtube-url> [style]"
----
-
-# YouTube to Infographic
-
-Transform YouTube video content into data visualization infographic.
-
-<youtube_url>$ARGUMENTS</youtube_url>
-
-## Workflow
-
-1. **Extract Video Data**
-   - Activate `youtube-handling` skill
-   - Run VidCap API:
-     ```bash
-     python .opencode/skills/youtube-handling/scripts/vidcap.py info "<url>"
-     python .opencode/skills/youtube-handling/scripts/vidcap.py summary "<url>" --screenshot
-     python .opencode/skills/youtube-handling/scripts/vidcap.py caption "<url>" --ext vtt
-     ```
-   - Extract: title, duration, key points, statistics, quotes
-
-2. **Analyze Content Structure**
-   - Identify infographic type based on content:
-     | Content Type | Chart Recommendation |
-     |--------------|---------------------|
-     | Steps/process | Timeline, Funnel |
-     | Comparisons | Bar Chart, Comparison Table |
-     | Statistics | KPI Cards, Icon Array |
-     | Breakdown | Pie/Donut, Treemap |
-     | Flow/journey | Sankey, Funnel |
-     | Rankings | Horizontal Bar |
-   - Activate `design-system` skill for chart selection
-
-3. **Design Infographic**
-   - Activate `ui-ux-pro-max` skill for visual style
-   - Inject brand context:
-     ```bash
-     node .opencode/skills/brand-guidelines/scripts/inject-brand-context.cjs --json
-     ```
-   - Get embeddable design tokens (MANDATORY for standalone HTML):
-     ```bash
-     node .opencode/skills/design-system/scripts/embed-tokens.cjs --minimal
-     ```
-   - Structure:
-     - Header: Video title + thumbnail
-     - Body: 3-5 key data visualizations
-     - Footer: Source attribution + CTA
-
-4. **Generate Visuals**
-   - Activate `ai-multimodal` skill
-   - Generate supporting images with Imagen 4 if needed
-   - Use brand colors from design tokens
-
-5. **Output**
-   - HTML infographic → `assets/infographic/{date}-{slug}.html`
-   - PNG export → `assets/infographic/{date}-{slug}.png`
-   - Source data → `assets/infographic/{date}-{slug}.json`
-
-## Skills Used
-
-- `youtube-handling` - VidCap API for video extraction
-- `design-system` - Chart type selection (slide-charts.csv)
-- `ui-ux-pro-max` - Visual design intelligence
-- `ai-multimodal` - Image generation with Imagen 4
-- `brand-guidelines` - Brand context injection
-- `assets-organizing` - Standardized output paths
-
-## Agents Used
-
-- `content-creator` - Content synthesis
-- `ui-ux-designer` - Visual layout
-
-## Infographic Styles
-
-| Style | Description | Best For |
-|-------|-------------|----------|
-| data | Statistics-heavy, chart-focused | Research, reports |
-| timeline | Chronological flow | How-to, processes |
-| comparison | Side-by-side analysis | Reviews, versus |
-| listicle | Numbered key points | Tips, summaries |
-| story | Narrative visual flow | Case studies |
-
-## Design Token Requirements (MANDATORY)
-
-All infographics MUST use design tokens from `assets/design-tokens.css`:
-
-| Element | Token Usage |
-|---------|-------------|
-| Background | `var(--color-background)`, `var(--primitive-gradient-dark)` |
-| Primary color | `var(--color-primary)` |
-| Secondary color | `var(--color-secondary)` |
-| Accent color | `var(--color-accent)` |
-| Text | `var(--color-foreground)`, `var(--color-foreground-secondary)` |
-| Cards | `var(--color-surface)`, `var(--card-radius)`, `var(--card-padding)` |
-| Font heading | `var(--typography-font-heading)` |
-| Font body | `var(--typography-font-body)` |
-| Spacing | `var(--primitive-spacing-*)` |
-| Font sizes | `var(--primitive-fontSize-*)` |
-| Gradients | `var(--primitive-gradient-primary)`, `var(--primitive-gradient-accent)` |
-
-**NEVER hardcode colors like `#e94560` or `#ff6b6b`** - always use CSS variables.
-
-## Output Format
-
-**IMPORTANT:** Embed tokens inline for file:// compatibility. Run:
-```bash
-node .opencode/skills/design-system/scripts/embed-tokens.cjs --minimal
-```
-
-```html
-<!DOCTYPE html>
-<html>
-<head>
-  <title>{Video Title} - Infographic</title>
-  <link href="https://fonts.googleapis.com/css2?family=Space+Grotesk:wght@400;500;600;700&family=Inter:wght@400;500;600&display=swap" rel="stylesheet">
-  <style>
-    /* Design Tokens - EMBED OUTPUT FROM embed-tokens.cjs */
-    :root {
-      /* Paste output from: node .opencode/skills/design-system/scripts/embed-tokens.cjs --minimal */
-    }
-  </style>
-  <style>
-    body {
-      font-family: var(--typography-font-body);
-      background: var(--color-background);
-      color: var(--color-foreground);
-    }
-    h1, h2, h3 {
-      font-family: var(--typography-font-heading);
-    }
-    /* All styles MUST use var(--token-name) */
-  </style>
-</head>
-<body>
-  <header><!-- Video title, thumbnail --></header>
-  <main><!-- Data visualizations --></main>
-  <footer><!-- Source: YouTube link --></footer>
-</body>
-</html>
-```
-
-## Quality Gate (Auto-Triggered)
-
-**After generating content, MUST run:**
-1. `/write/audit` → Score content (target: ≥8.0)
-2. `/write/publish` → Auto-fix if score <8.0
-3. Present final version with score
-
-## Examples
-
-```bash
-/youtube:infographic "https://youtube.com/watch?v=xyz"
-/youtube:infographic "https://youtu.be/abc" timeline
-/youtube:infographic "https://youtube.com/watch?v=123" data
-```

package/.opencode/commands/ckm-youtube-social.md

@@ -1,169 +0,0 @@
----
-description: "Convert YouTube video to multi-platform social posts - Args: <youtube-url> [platforms]"
----
-
-# YouTube to Social Posts
-
-Transform YouTube video into platform-optimized social content.
-
-<youtube_url>$ARGUMENTS</youtube_url>
-
-## Workflow
-
-1. **Extract Video Highlights**
-   - Activate `youtube-handling` skill
-   - Run VidCap API:
-     ```bash
-     # Get summary with key points
-     python .opencode/skills/youtube-handling/scripts/vidcap.py summary "<url>" --screenshot
-
-     # Get comments for social proof angles
-     python .opencode/skills/youtube-handling/scripts/vidcap.py comments "<url>" --order relevance
-
-     # Get video info for metadata
-     python .opencode/skills/youtube-handling/scripts/vidcap.py info "<url>"
-     ```
-   - Extract: title, key points, quotable moments, statistics, thumbnail
-
-2. **Generate Platform Content**
-   - Activate `social-media` skill
-   - Activate `copywriting` skill for hooks
-   - Load `references/hook-writing.md` for hook patterns
-   - Create content per platform:
-
-   | Platform | Format | Length | Hook Style |
-   |----------|--------|--------|------------|
-   | Twitter/X | Thread or single | 280 chars | Curiosity gap |
-   | LinkedIn | Post + carousel | 3000 chars | Thought leadership |
-   | Instagram | Caption + carousel | 2200 chars | Story opener |
-   | TikTok | Script for video | 60-90 sec | 3-second hook |
-   | YouTube | Community post | 500 chars | Engagement question |
-
-3. **Hook Writing Rules**
-   - Reference `social-media/references/hook-writing.md`
-   - AVOID:
-     - Punching keywords ("SHOCKING!", "INSANE!")
-     - Clickbait ("You won't BELIEVE...")
-     - Excessive emojis
-     - Empty promises
-   - USE:
-     - Curiosity gaps
-     - Contrarian takes
-     - Story openers
-     - Specific numbers
-
-4. **Visual Assets**
-   - Extract key frames from video screenshots
-   - Generate carousel graphics if needed (ai-multimodal)
-   - Apply brand styling
-
-5. **Output Structure**
-   - Per platform file → `assets/posts/{platform}/{date}-{slug}.md`
-   - Combined file → `assets/posts/{date}-{slug}-all-platforms.md`
-
-## Skills Used
-
-- `youtube-handling` - VidCap API for video extraction
-- `social-media` - Platform specs, posting best practices
-- `copywriting` - Hook writing, CTAs
-- `brand-guidelines` - Brand voice consistency
-- `assets-organizing` - Standardized output paths
-
-## Agents Used
-
-- `social-media-manager` - Multi-platform content
-- `copywriter` - Engaging hooks
-
-## Platform Specifications
-
-### Twitter/X
-- Max: 280 characters (threads: 25 tweets)
-- Image: 1200x675 (16:9)
-- Hook: Open loop, question, contrarian
-
-### LinkedIn
-- Max: 3000 characters
-- Image: 1200x627 or 1080x1080
-- Hook: Professional insight, data point
-
-### Instagram
-- Caption: 2200 characters
-- Carousel: 10 slides, 1080x1080 or 1080x1350
-- Hook: Story, question, bold statement
-
-### TikTok
-- Script: 60-90 seconds
-- Hook: First 3 seconds critical
-- Style: Authentic, conversational
-
-### YouTube Community
-- Max: 500 characters
-- Engagement: Polls, questions
-- CTA: Watch full video
-
-## Output Format
-
-```markdown
-# Social Posts: {Video Title}
-
-Source: {youtube_url}
-
----
-
-## Twitter/X
-
-### Thread
-1/ {Hook - curiosity gap}
-
-{Key point}
-
-2/ {Insight}
-
-{Supporting detail}
-
-...
-
-{n}/ Full video: {link}
-
----
-
-## LinkedIn
-
-{Professional hook with data point}
-
-{3-5 paragraphs of value}
-
-What's your take on {topic}?
-
-#relevant #hashtags
-
----
-
-## Instagram
-
-{Story-driven hook}
-
-{Value points with emoji bullets}
-
-Save this for later!
-
-.
-.
-.
-#hashtags
-```
-
-## Quality Gate (Auto-Triggered)
-
-**After generating content, MUST run:**
-1. `/write/audit` → Score content (target: ≥8.0)
-2. `/write/publish` → Auto-fix if score <8.0
-3. Present final version with score
-
-## Examples
-
-```bash
-/youtube:social "https://youtube.com/watch?v=xyz"
-/youtube:social "https://youtu.be/abc" twitter,linkedin
-/youtube:social "https://youtube.com/watch?v=123" instagram
-```

package/.opencode/package.json

@@ -1,8 +0,0 @@
-{
-  "name": "@claudekit/opencode-plugins",
-  "version": "1.0.0",
-  "description": "ClaudeKit hooks converted to OpenCode plugins",
-  "dependencies": {
-    "@opencode-ai/plugin": ">=0.1.0"
-  }
-}

package/.opencode/plugin/context-injector.ts

@@ -1,84 +0,0 @@
-import type { Plugin } from "@opencode-ai/plugin";
-
-const { buildReminderContext } = require("./lib/context-builder.cjs");
-const { detectProject, getCodingLevelGuidelines } = require("./lib/project-detector.cjs");
-const { loadConfig } = require("./lib/ck-config-utils.cjs");
-
-// Track first message per session to inject context once
-const injectedSessions = new Set<string>();
-
-/**
- * Context Injector Plugin - Inject session context into first message
- *
- * Combines functionality of dev-rules-reminder.cjs and session-init.cjs.
- * Injects rules, session info, project detection into first user message only.
- */
-export const ContextInjectorPlugin: Plugin = async ({ directory }) => {
-  // Load config once at plugin initialization
-  let config: any;
-  let detections: any;
-
-  try {
-    config = loadConfig();
-    detections = detectProject();
-  } catch (e) {
-    // Fallback to defaults if config loading fails
-    config = { codingLevel: -1 };
-    detections = {};
-  }
-
-  return {
-    "chat.message": async ({}: any, { message }: any) => {
-      // Get or generate session ID
-      const sessionId = process.env.OPENCODE_SESSION_ID ||
-                        `opencode-${Date.now()}`;
-
-      // Only inject on first message per session
-      if (injectedSessions.has(sessionId)) {
-        return;
-      }
-      injectedSessions.add(sessionId);
-
-      try {
-        // Build context
-        const { content } = buildReminderContext({
-          sessionId,
-          config,
-          staticEnv: {
-            nodeVersion: process.version,
-            osPlatform: process.platform,
-            gitBranch: detections.gitBranch,
-            gitRoot: detections.gitRoot,
-            user: process.env.USER || process.env.USERNAME,
-            locale: process.env.LANG || '',
-            timezone: Intl.DateTimeFormat().resolvedOptions().timeZone
-          },
-          configDirName: '.opencode'
-        });
-
-        // Inject coding level guidelines if configured
-        const codingLevel = config.codingLevel ?? -1;
-        const guidelines = getCodingLevelGuidelines(codingLevel, `${directory}/.opencode`);
-
-        // Prepend context to first user message
-        const contextBlock = [
-          '<system-context>',
-          content,
-          guidelines ? `\n${guidelines}` : '',
-          '</system-context>',
-          ''
-        ].filter(Boolean).join('\n');
-
-        // Modify message content (prepend context)
-        if (message && typeof message.content === 'string') {
-          message.content = contextBlock + message.content;
-        }
-      } catch (e) {
-        // Silently fail - don't break the chat if context injection fails
-        console.error('[ContextInjector] Failed to inject context:', e);
-      }
-    }
-  };
-};
-
-export default ContextInjectorPlugin;