Docs

Agent & Search

The /agent and /search commands represent Agentastic's most advanced AI capabilities, enabling autonomous task completion and intelligent information retrieval.

Agent Command (/agent)

The /agent command activates an autonomous AI assistant that can break down complex tasks, execute multiple steps, and deliver comprehensive results.

What Makes /agent Different

Unlike single-response AI commands, /agent:

  • Plans multi-step approaches
  • Executes tasks autonomously
  • Adapts based on results
  • Combines multiple tools
  • Iterates until complete

Agent Capabilities

Research and Analysis

/agent Research competitors' pricing and create comparison table
/agent Analyze customer feedback and identify top 3 issues
/agent Study best practices for React performance optimization

Content Creation

/agent Write comprehensive guide on macOS productivity
/agent Create weekly newsletter from recent blog posts
/agent Generate documentation from code comments

Task Automation

/agent Organize my Downloads folder by type and date
/agent Find and summarize all meeting notes from this week
/agent Create project status report from multiple sources

Problem Solving

/agent Debug why the application is running slowly
/agent Find solution for the Excel formula error
/agent Optimize this workflow for efficiency

How /agent Works

  1. Task Analysis - Breaks down your request
  2. Planning - Creates step-by-step approach
  3. Execution - Performs each step
  4. Validation - Checks results
  5. Iteration - Refines if needed
  6. Delivery - Presents final outcome

Agent Examples

Complex Research

/agent Research and summarize the latest developments in AI text generation, including key players, recent breakthroughs, and future predictions. Create a structured report with sources.

Multi-Source Analysis

/agent Analyze my email, calendar, and recent documents to create a comprehensive weekly status update for my manager.

Workflow Automation

/agent Every Friday, check my calendar for next week, identify preparation needed for each meeting, and create a preparation checklist.

Agent Best Practices

Be Specific About Outcomes

/agent Help with project/agent Create project timeline with milestones based on requirements.doc

Provide Context

/agent Fix the problem/agent Debug why users can't log in after the recent update

Set Clear Boundaries

/agent Do everything needed/agent Research, outline, and draft introduction for blog post

Agent Limitations

  • Requires clear task definition
  • May take longer for complex tasks
  • Internet connection required
  • Some actions need confirmation
  • Resource intensive for large tasks

The /search command provides intelligent web search with understanding of context and intent.

Search Capabilities

Intelligent Query Understanding

/search Latest macOS shortcuts for productivity
/search How to fix Python ImportError
/search Reviews for standing desks under $500

Multi-Source Aggregation

/search Compare iPhone 15 Pro vs Samsung S24 Ultra
/search Collect recipes for vegetarian lasagna
/search Find all tutorials on React Server Components

Real-Time Information

/search Current stock price of AAPL
/search Weather forecast for next week
/search Latest news about artificial intelligence

Advanced Search Features

Filtered Searches

/search:academic Papers on machine learning bias
/search:news Recent developments in space technology
/search:tutorials How to use Docker for beginners

Time-Bounded Searches

/search:recent Last 24 hours of tech news
/search:week Updates from this week on project
/search:month Monthly summary of AI developments

Domain-Specific

/search site:github.com React best practices
/search site:stackoverflow.com Python async await
/search site:reddit.com Reviews of productivity apps

Search Intelligence

The search command goes beyond simple queries:

  1. Intent Recognition - Understands what you're looking for
  2. Source Evaluation - Prioritizes reliable sources
  3. Result Synthesis - Combines multiple sources
  4. Fact Checking - Validates information
  5. Summary Generation - Distills key points

Search Examples

Technical Problem Solving

/search Error: "Module not found" in Next.js 13
Result: Targeted solutions from Stack Overflow, GitHub issues, and official docs

Product Research

/search Best ergonomic keyboard for programming 2024
Result: Comparison of top options with pros, cons, and user reviews

Learning Resources

/search Free courses on machine learning for beginners
Result: Curated list of courses with difficulty levels and time requirements

Use both commands together for powerful workflows:

Research Projects

/agent /search Create comprehensive report on remote work trends post-2023

Problem Investigation

/search Find similar issues → /agent Analyze and create solution plan

Content Creation

/search Gather sources → /agent Write article with citations

Agent vs Search: When to Use Which

Use /agent When:

  • Task requires multiple steps
  • Need analysis and synthesis
  • Want autonomous execution
  • Dealing with complex problems
  • Creating comprehensive output

Use /search When:

  • Need specific information
  • Want quick answers
  • Looking for current data
  • Researching options
  • Fact-checking claims

Advanced Techniques

Agent Chaining

/agent Create outline → Review → /agent Expand each section

Search Refinement

/search Initial query → Too broad → /search:specific Refined query

Hybrid Workflows

/search Gather data → /agent Analyze → /search Verify findings

Performance Optimization

For /agent

  • Break very large tasks into phases
  • Provide clear success criteria
  • Include relevant context upfront
  • Use specific output formats
  • Use precise keywords
  • Specify domains when known
  • Add time constraints if relevant
  • Request summaries for long results

Common Use Cases

Daily Productivity

/agent Summarize important emails and create action items
/search Quick answers to technical questions
/agent Prepare for tomorrow's meetings

Research and Learning

/search Find learning resources on new topic
/agent Create study guide from multiple sources
/search Stay updated on industry trends

Problem Solving

/search Find others with similar issues
/agent Develop solution strategy
/search Verify solution effectiveness

Troubleshooting

Agent Taking Too Long

  • Simplify the task
  • Break into smaller parts
  • Check internet connection
  • Cancel and restart if needed

Search Not Finding Results

  • Try different keywords
  • Remove filters
  • Check spelling
  • Broaden search terms

Quality Issues

  • Provide more context
  • Specify desired format
  • Request specific sources
  • Iterate on results

Tips and Tricks

Agent Mastery

  1. Start with simple tasks to understand capabilities
  2. Gradually increase complexity
  3. Save successful agent prompts
  4. Learn from agent's approach

Search Excellence

  1. Use natural language queries
  2. Combine with other tools
  3. Leverage search history
  4. Refine based on results

Next Steps

The /agent and /search commands represent the pinnacle of Agentastic's AI capabilities. Master these to transform how you work with information and complete complex tasks!