### Run SkillPointer Setup Script (Bash)

Source: https://context7.com/blacksiders/skillpointer/llms.txt

Executes the main Python setup script to categorize skills and generate category pointers. This script analyzes skill names using keyword heuristics and migrates raw skills to a hidden vault, then creates lightweight pointers for dynamic retrieval.

```bash
python setup.py
```

--------------------------------

### Windows Installation Scripts (Batch/VBS)

Source: https://context7.com/blacksiders/skillpointer/llms.txt

Provides batch and VBScript files for installing SkillPointer on Windows systems. The batch file displays console output during installation, while the VBScript performs a silent installation with a success dialog.

```batch
REM Using the batch file (shows console output)
Install.bat

REM Using the VBS file (silent installation with success dialog)
cscript Install.vbs
```

--------------------------------

### Manual Implementation: Creating Custom Pointers

Source: https://context7.com/blacksiders/skillpointer/llms.txt

This bash script demonstrates the manual steps to create a custom category pointer. It involves setting up the hidden vault structure, moving a custom skill, creating a pointer directory, and defining the pointer's SKILL.md file.

```bash
# Step 1: Create hidden vault structure
mkdir -p ~/.opencode-skill-libraries/custom-category
mv ~/.config/opencode/skills/my-custom-skill ~/.opencode-skill-libraries/custom-category/

# Step 2: Create pointer directory
mkdir -p ~/.config/opencode/skills/custom-category-pointer

# Step 3: Create pointer SKILL.md
cat > ~/.config/opencode/skills/custom-category-pointer/SKILL.md << 'EOF'
---
name: custom-category-pointer
description: Triggers for custom domain tasks. Pointer to specialized skill library.
---

# Custom Category Library 🎯

Use your file tools to browse: `~/.opencode-skill-libraries/custom-category`

1. Use `list_dir` to see available skills
2. Use `view_file` to read the specific SKILL.md you need
3. Follow the instructions in that skill file

**Library Path:** `~/.opencode-skill-libraries/custom-category`
EOF

```

--------------------------------

### AI Execution Flow for Category Pointer Usage

Source: https://context7.com/blacksiders/skillpointer/llms.txt

This illustrates the step-by-step process an AI agent follows when a user prompt triggers a category pointer. It shows how the AI navigates through different levels of skill retrieval, from matching the pointer to executing a specific skill.

```text
User Prompt:
"I want to create a CSS button. Please consult your `web-dev-category-pointer`
first to find the exact best practice from your library before writing the code."

AI Execution Flow:
1. AI matches "web-dev-category-pointer" from <available_skills> (Level 1)
2. AI loads the pointer's SKILL.md body (Level 2 - just ~150 tokens)
3. AI follows instructions, runs: list_dir("~/.opencode-skill-libraries/web-dev")
4. AI identifies relevant skill: "css-button-best-practices"
5. AI reads: view_file("~/.opencode-skill-libraries/web-dev/css-button-best-practices/SKILL.md")
6. AI generates expert-level code based on the specific skill content

```

--------------------------------

### Skill Migration Directory Structure (Conceptual)

Source: https://context7.com/blacksiders/skillpointer/llms.txt

Illustrates the change in directory structure before and after applying SkillPointer. Skills are moved from the active scan path to a hidden vault (`~/.opencode-skill-libraries/`) and organized by category.

```text
# Directory structure before SkillPointer:
# ~/.config/opencode/skills/
# ├── react-best-practices/SKILL.md
# ├── aws-lambda-guide/SKILL.md
# ├── pytorch-training-tips/SKILL.md
# └── ... (2000+ skill folders)

# Directory structure after SkillPointer:
# ~/.config/opencode/skills/
# ├── web-dev-category-pointer/SKILL.md      # ~35 lightweight pointers
# ├── devops-category-pointer/SKILL.md
# ├── ai-ml-category-pointer/SKILL.md
# └── ...
#
# ~/.opencode-skill-libraries/               # Hidden vault (not scanned)
# ├── web-dev/
# │   └── react-best-practices/SKILL.md
# ├── devops/
# │   └── aws-lambda-guide/SKILL.md
# ├── ai-ml/
# │   └── pytorch-training-tips/SKILL.md
# └── ... (all 2000+ skills organized by category)

```

--------------------------------

### Web Development Category Pointer SKILL.md Template

Source: https://context7.com/blacksiders/skillpointer/llms.txt

This markdown file serves as a template for a web development category pointer. It instructs the AI to use file reading tools to access a hidden library of specialized web development skills.

```markdown
---
name: web-dev-category-pointer
description: Triggers when encountering any task related to web-dev. This is a pointer to a library of specialized skills.
---

# Web Dev Capability Library 🎯

You do not have all Web Dev skills loaded immediately in your background context.
Instead, you have access to a rich library of 156 highly-specialized skills on
your local filesystem.

## Instructions
1. When you need to perform a task related to web-dev, you MUST use your file
   reading tools (like `list_dir` and `view_file` or `read_file`) to browse
   the hidden library directory: `/home/user/.opencode-skill-libraries/web-dev`
2. Locate the specific Markdown files related to the exact sub-task you need.
3. Read the relevant Markdown file(s) into your context.
4. Follow the specific instructions and best practices found within those files
   to complete the user's request.

## Available Knowledge
This library contains 156 specialized skills covering various aspects of Web Dev.

**Hidden Library Path:** `/home/user/.opencode-skill-libraries/web-dev`

*Reminder: Do not guess best practices or blindly search GitHub. Always consult
your local library files first.*

```

--------------------------------

### Skill Categorization Heuristics (Python)

Source: https://context7.com/blacksiders/skillpointer/llms.txt

Defines a dictionary of domain heuristics used to categorize skills based on keyword matching. The `get_category_for_skill()` function (not shown) would use this mapping to assign skills to categories like 'security', 'web-dev', 'devops', etc.

```python
# The heuristic engine maps skills to categories based on name patterns
DOMAIN_HEURISTICS = {
    "security": ["attack", "injection", "vulnerability", "xss", "penetration",
                 "fuzzing", "auth", "jwt", "oauth", "exploit", "encryption"],
    "web-dev": ["angular", "react", "vue", "tailwind", "frontend", "css",
                "html", "nextjs", "svelte", "astro", "web", "dom"],
    "devops": ["aws", "azure", "docker", "kubernetes", "ci-cd", "terraform",
               "ansible", "github-actions", "jenkins", "cloud", "linux"],
    "ai-ml": ["ai-", "ml-", "llm", "agent", "gpt", "claude", "openai",
              "anthropic", "prompt", "rag", "pytorch", "tensorflow"],
    "database": ["sql", "mysql", "postgres", "mongo", "redis", "prisma",
                 "orm", "nosql", "supabase", "sqlite"],
    # ... 25+ more categories including: backend-dev, mobile, game-dev,
    # blockchain, data-science, testing, automation, design, etc.
}

# Example categorization results:
# "react-best-practices"     → web-dev
# "aws-lambda-guide"         → devops
# "pytorch-training-tips"    → ai-ml
# "sql-injection-prevention" → security
# "unknown-random-skill"     → _uncategorized

```

=== COMPLETE CONTENT === This response contains all available snippets from this library. No additional content exists. Do not make further requests.