Analyze entity-level code changes with tree-sitter and Git.
npx skills add https://github.com/joegoldin/agent-skills --skill sem-joegoldin
Please help me install this Agent Skill. Skill: sem Source: https://github.com/joegoldin/agent-skills/tree/main/skills/sem Command: npx skills add https://github.com/joegoldin/agent-skills --skill sem-joegoldin
This Skill provides a detailed view of code changes at the entity level, enabling precise understanding and impact analysis without delving into line-level diffs.
sem impact to see what depends on the entity you're modifying, and sem context to get a token-budgeted LLM context for the symbol.Analyze the impact of changes on the 'authenticateUser' function with sem impact authenticateUser.
High-intent search queries and answers about installing and using this skill.
Entity-level code analysis parses your repository with tree-sitter to detect and match functions, methods, and classes across versions. This approach provides a structured view of modifications, letting you assess changes without reading raw line-level diffs.
To assess the impact of modifying a function in a Git repository, you can fetch its dependencies and dependents. This reveals what relies on the entity you are changing, helping you understand the potential effects before refactoring.
Yes, entity-level diffing detects renamed and moved functions across different versions. By using structured parsing, it tracks entities through changes, ensuring accurate attribution of modifications in your Git history.
Yes, you need tree-sitter installed to perform entity-level code analysis. The tool relies on tree-sitter for structured parsing, and you must have a Git environment configured to analyze the repository.
You can generate token-budgeted LLM context for a specific code symbol by requesting contextual information for that entity. This fetches relevant dependencies and dependents to provide an informed, size-limited context window.
xberg-ioGenerate syntax trees from source code using tree-sitter.
zot24Provide persistent memory and context retrieval for LLM applications.
dyCuong03Snapshot entity IDs and component data into NativeArrays for safe structural changes.
kwakseongjaeAdapt Korean text to English, Japanese, and Traditional Chinese with cultural accuracy.