Analyze File or Directory with Ollama

Analyze the specified file or directory using the ollama agent pipeline.

Usage: /analyze <file_or_directory> [focus_area]

Examples:

• /analyze src/auth.py security - Security analysis of single file
• /analyze README.md - General analysis
• /analyze src/ architecture - Architecture analysis of directory
• /analyze src/ security - Security audit of entire directory

You are an intelligent task router for ollama-based analysis.

Task: Analyze the target at path: $1 Focus Area: $2

Your Process:

1. Check Target:

   • Determine if target is file or directory
   • For files: Get size and estimate tokens
   • For directories: Use directory operations for efficiency

2. Select Strategy:

   For Files:

   • Small files (< 20KB): Direct ollama-prompt
   • Large files (> 20KB): Use ollama-chunked-analyzer approach

   For Directories:

   • Architecture focus: Use @./dir/:tree for structure analysis
   • Security focus: Use @./dir/:search:PATTERN for vulnerability patterns
   • Quality focus: Use @./dir/:search:TODO + @./dir/:tree
   • General: Use @./dir/:tree then targeted analysis

3. Directory Operations by Focus Area:

   Focus Area	Primary Operation	Example
   security	@./dir/:search:	@./src/:search:eval, @./src/:search:password
   architecture	@./dir/:tree	@./src/:tree
   performance	@./dir/:search:	@./src/:search:for.*in, @./src/:search:query
   quality	@./dir/:search: + :tree	@./src/:search:TODO, @./src/:tree
   general	@./dir/:tree	@./src/:tree

4. Invoke Agent: Use the Task tool to invoke the ollama-task-router agent:

   • Pass the target path: $1
   • Pass the focus area: $2
   • Agent uses appropriate directory operations
   • Agent handles model selection and execution

5. Agent Will:

   • Detect if target is file or directory
   • Select appropriate model (kimi-k2-thinking, deepseek, qwen3-vl)
   • Use directory operations for directories
   • Route to chunked analyzer if needed
   • Execute analysis with ollama-prompt
   • Return synthesized results

6. Your Role:

   • Receive agent's analysis report
   • Present findings to user concisely
   • Highlight critical issues
   • Provide actionable recommendations

Focus Areas:

• security: Vulnerabilities, attack vectors, security best practices
• architecture: Design patterns, scalability, maintainability
• performance: Bottlenecks, optimization opportunities
• quality: Code quality, best practices, refactoring needs
• general: Comprehensive overview

Directory Analysis Benefits:

• @./dir/:tree uses ~500 tokens vs ~15,000 for reading files individually
• @./dir/:search:PATTERN finds specific issues across entire codebase
• More efficient and comprehensive than file-by-file analysis

Remember: This delegates to ollama to save your context budget!