Available Tools

Explore all the powerful tools available to your AI assistant

Focused Toolset Philosophy

We maintain a deliberately tight set of tools, focusing on kit's unique strengths rather than duplicating what AI assistants already do well. We enhance existing tools with new features instead of adding redundant ones. This keeps context clean and responses fast.

Documentation Research

deep_research_package

Get comprehensive package documentation using LLM knowledge

Natural Language

"Using Kit, research the documentation for React hooks"

Parameters

package_name

query

Function Signature

deep_research_package({
  "package_name": "react",
  "query": "How do hooks work?"  // optional specific question
})

Repository Management

open_repository

Open a local directory or GitHub repository for analysis

Natural Language

"Using Kit, open the repository at /path/to/project"

Parameters

path_or_url

github_token

ref

Function Signature

open_repository({
  "path_or_url": "/Users/me/project",
  "github_token": "ghp_...",  // optional
  "ref": "main"  // optional
})

get_file_tree

Get the complete file structure of the repository

Natural Language

"Using Kit, show me the file tree of this repository"

Parameters

repo_id

Function Signature

get_file_tree({
  "repo_id": "repo_123"
})

Code Intelligence

extract_symbols

Extract functions, classes, and other symbols from code

Natural Language

"Using Kit, extract all symbols from the main.py file"

Parameters

repo_id

file_path

symbol_type

Function Signature

extract_symbols({
  "repo_id": "repo_123",
  "file_path": "src/main.py",
  "symbol_type": "function"  // optional: function, class
})

find_symbol_usages

Find where a symbol is used throughout the codebase

Natural Language

"Using Kit, find all usages of the Database class"

Parameters

repo_id

symbol_name

file_path

Function Signature

find_symbol_usages({
  "repo_id": "repo_123",
  "symbol_name": "Database",
  "file_path": "src/"  // optional: narrow search
})

Code Search

grep_code

Advanced regex search with context lines

Natural Language

"Using Kit, grep for import statements with context"

Parameters

repo_id

pattern

path

context_lines

Function Signature

grep_code({
  "repo_id": "repo_123",
  "pattern": "^import .*",
  "path": "src/",
  "context_lines": 2
})

grep_ast

Search code using AST patterns for semantic code structure matching

Natural Language

"Using Kit, find all async functions in the codebase"

Parameters

repo_id

pattern

mode

file_pattern

max_results

Function Signature

grep_ast({
  "repo_id": "repo_123",
  "pattern": "async def",  // or '{"type": "try_statement"}'
  "mode": "simple",  // simple, pattern, or query
  "file_pattern": "**/*.py",
  "max_results": 20
})

Git Information

review_diff

Review code changes with AI-powered analysis

Natural Language

"Using Kit, review the diff between main and feature branch"

Parameters

repo_id

diff_spec

priority_filter

max_files

model

Function Signature

review_diff({
  "repo_id": "repo_123",
  "diff_spec": "main..feature-branch",
  "priority_filter": ["high", "medium"],  // optional
  "max_files": 10,  // optional limit
  "model": "gpt-4"  // optional
})