Deep Research Sub-Agent

VerifiedSafe

Launches a sub-agent to conduct in-depth research on a given topic or codebase question. It follows a structured process: defining scope, analyzing the codebase with the Explore sub-agent, consulting external sources, and synthesizing findings into a comprehensive report. Useful for investigating complex issues or unfamiliar code quickly.

Sby Skills Guide Bot
DevelopmentIntermediate
806/2/2026
Claude Code
#deep-research#codebase-analysis#external-research#synthesis

Recommended for

Our review

Launches a sub-agent to conduct comprehensive research on a specified topic or codebase question.

Strengths

  • Thorough investigation covering codebase analysis, external research, and synthesis.
  • Structured process that ensures all aspects are explored.
  • Produces a detailed report with summary, analysis, file locations, recommendations, and open questions.

Limitations

  • May require multiple steps and sub-agent calls, increasing time and resource usage.
  • Dependent on the quality of the Explore sub-agent and external sources.
  • Not suitable for extremely broad or vague topics without clear scope.
When to use it

Use when you need an in-depth understanding of a complex topic, code pattern, or architectural decision.

When not to use it

Avoid for simple lookups or when a quick answer from existing knowledge is sufficient.

Security analysis

Safe
Quality score80/100

The skill provides research instructions for a sub-agent without declaring or using any tools. It does not contain destructive commands, exfiltration patterns, or obfuscated payloads. No execution risk is present.

No concerns found

Examples

Research authentication flow
Perform deep research on the authentication flow in our codebase, including the login, token refresh, and permission checks. Provide a detailed report.
Best practices for error handling
Research best practices for error handling in Python. Find external resources and analyze our current error handling patterns in the src/ directory.
Architecture of the payment module
Deep research on the architecture of the payment module. Understand the dependencies, data flow, and key design decisions.

Deep Research Sub-Agent

Launch a sub-agent to perform deep research on a topic or codebase question.

Topic: $ARGUMENTS

Instructions for Sub-Agent

You are a research specialist. Your goal is to thoroughly investigate the given topic and provide comprehensive findings.

Research Process:

  1. Scope Definition

    • Clarify what needs to be researched
    • Identify key questions to answer
  2. Codebase Analysis (if code-related)

    • Use Explore sub-agent to search relevant files
    • Trace code paths and dependencies
    • Understand architecture and patterns
  3. External Research (if needed)

    • Search documentation
    • Look up best practices
    • Find relevant examples
  4. Synthesis

    • Compile findings
    • Identify patterns and insights
    • Note any gaps or uncertainties

Output:

Provide a research report with:

  • ๐Ÿ“‹ Summary of findings
  • ๐Ÿ” Detailed analysis
  • ๐Ÿ“ Relevant files/code locations
  • ๐Ÿ’ก Recommendations or next steps
  • โ“ Open questions (if any)

Use Explore sub-agent (with Sonnet for complex analysis) to investigate thoroughly.

Related skills