How do I ensure the skill has the data it needs to work?

You must verify that claudemem v0.2.0 is installed and that you have run the indexing process with enrichment enabled (claudemem index --enrich) to generate the necessary file and symbol summaries.

Why is grep forbidden when using the Ultrathink Detective skill?

This skill relies on semantic understanding rather than simple text matching. By using claudemem v0.2.0, it can identify behavior, intent, and side effects that traditional grep or find tools would overlook.

Which AI model is used for these investigations?

Ultrathink Detective is specifically configured to use the Opus model, which provides the high-level reasoning and multi-perspective thinking required for deep codebase analysis.

What is the benefit of the claudemem v0.2.0 enrichment layer?

The enrichment layer uses an LLM to generate file and symbol summaries. This allows the detective skill to understand the 'purpose' of a file and the 'behavior' of a function beyond just the raw code.

Ultrathink Detective

Name: Ultrathink Detective
Author: MadAppGang

byMadAppGang

•

177

Security & Testing

Performs deep, multi-dimensional codebase audits and architectural analysis using LLM-enriched semantic memory.

About

Ultrathink Detective is a high-level analysis skill designed for senior engineers and tech leads. It leverages the Opus model and claudemem v0.2.0 to provide a holistic view of codebases by combining file-level architectural summaries, symbol-level behavioral analysis, and raw code inspection. By strictly prioritizing semantic vector search over traditional text-matching tools like grep, it detects complex patterns, hidden side effects, and architectural gaps that standard tools miss, making it ideal for large-scale refactoring and security audits.

Key Features

Comprehensive auditing for security, performance, reliability, and technical debt
Multi-perspective analysis combining architect, developer, tester, and debugger viewpoints
LLM-enriched semantic indexing using file_summary and symbol_summary layers
Advanced pattern recognition across complex, multi-layered software architectures
Evidence-based reporting with direct links to semantic code chunks
177 GitHub stars

Use Cases

Root cause analysis for complex bugs spanning multiple decoupled systems
Comprehensive security audits and input validation vulnerability scanning
Pre-refactoring technical debt assessment and architectural decision mapping

About

Key Features

Comprehensive auditing for security, performance, reliability, and technical debt
Multi-perspective analysis combining architect, developer, tester, and debugger viewpoints
LLM-enriched semantic indexing using file_summary and symbol_summary layers
Advanced pattern recognition across complex, multi-layered software architectures
Evidence-based reporting with direct links to semantic code chunks
177 GitHub stars

Use Cases

Root cause analysis for complex bugs spanning multiple decoupled systems
Comprehensive security audits and input validation vulnerability scanning
Pre-refactoring technical debt assessment and architectural decision mapping