investigate-dependencies

The investigate-dependencies skill is a specialized engineering utility designed for codebase maintenance, architectural optimization, and technical debt reduction. This skill enables developers and agents to systematically audit third-party libraries and internal modules to ensure that the repository remains lean and performant. By conducting a line-by-line dependency walkthrough, it detects instances where custom code reimplements logic that is already natively supported by imported frameworks or standard libraries, such as Pydantic for validation, standard datetime libraries, or specialized utility modules. This is essential for projects using high-overhead frameworks like FastAPI, SQLAlchemy, or complex machine learning stacks where improper usage patterns can lead to token bloat or unnecessary computational costs.

• Performs comprehensive import inventory creation to map dependencies against documentation sources.

• Executes line-by-line dependency usage examination to uncover misuse, missing error handling, or abandoned library features.

• Automates redundancy detection in critical areas including datetime operations, JSON handling, data validation, and configuration management.

• Conducts capability gap analysis to identify unused library features that could simplify code or replace custom, fragile implementations.

• Generates structured Markdown reports including executive summaries, inventory tables, and prioritized refactoring recommendations.

• Targets engineering teams looking to reduce codebase complexity, minimize dependencies, and enforce consistent usage of standardized utilities.

• The skill requires a clear target file or module scope as input; it outputs a detailed dependency investigation report.

• Practical usage involves running the protocol during code reviews or when refactoring legacy modules to modernize them with current project best practices.

• Users should monitor for potential 'over-optimization' where replacing simple custom code with heavy dependencies might introduce unnecessary bloat or maintenance overhead.

• Constraints include the need for accurate documentation links and the agent's ability to interpret library source code for advanced capability discovery.

• Ideal for environments using Python-based ecosystems where managing Pydantic models, environment configurations, and external API integrations is constant.