From kitty
Structural codebase exploration using Cartographing Kittens' graph. Use when the user asks "what's in this file", "how is this organized", "explore the codebase", "navigate the code", "understand the structure", "what does this module contain", or wants to browse definitions, imports, and relationships. Prefer over grep/glob for structural and relational questions about code organization.
How this skill is triggered — by the user, by Claude, or both
Slash command
/kitty:kitty-exploreThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
Explore codebase structure through the Cartographing Kittens graph — definitions, imports,
Explore codebase structure through the Cartographing Kittens graph — definitions, imports, call relationships, and inheritance.
The orchestrator builds all context directly — no agent dispatch needed for exploration. Agents CANNOT call MCP tools, so the orchestrator IS the explorer.
index_codebase(full=false) to ensure the graph is freshannotation_status() — check coverage percentage
kitty:annotate first."Broad overview (file or module exploration):
get_context_summary(file_paths=[...]) for a token-efficient overview of files/modulesrank_nodes(scope=target_files, limit=10) to surface the most important symbolsMulti-symbol deep dive:
batch_query_nodes(names=[...], include_neighbors=true) for detailed info on multiple symbols at onceRelationship exploration:
find_dependents(name=..., max_depth=2) for "who uses this?"find_dependencies(name=..., max_depth=2) for "what does this depend on?"Semantic/concept search:
search(query="keyword") (best after annotation)Single symbol lookup:
query_node(name="ClassName") for a single symbol with neighborsFormat the gathered context as structured output for the user. Include:
rank_nodes)| Tool | Use for |
|---|---|
index_codebase | Ensure the graph is up to date |
annotation_status | Check annotation coverage, warn if low |
get_context_summary | Token-efficient overview of files/modules |
batch_query_nodes | Look up multiple symbols at once with neighbors |
get_file_structure | See all definitions in a single file |
query_node | Look up a single symbol with neighbors |
find_dependents | Who uses this symbol? (blast radius) |
find_dependencies | What does this symbol depend on? |
rank_nodes | Surface the most important symbols in a scope |
search | Find nodes by name or summary (best after annotation) |
:: separator: module.path::ClassName::method_namemodule, class, function, method, variableimports, calls, inherits, contains, depends_onget_context_summary for broad exploration — it is more token-efficient than calling get_file_structure + query_node for each filebatch_query_nodes instead of multiple query_node calls when exploring several symbolsrank_nodes to identify the most structurally important symbols before diving deeperannotation_status first — if many
nodes are pending, run kitty:annotate before searchingreferences/tool-reference.md for full parameter detailsnpx claudepluginhub kakise/cartographing-kitties-plugin --plugin kittyProvides behavioral guidelines to reduce common LLM coding mistakes, focusing on simplicity, surgical changes, assumption surfacing, and verifiable success criteria.
Searches, retrieves, and installs Agent Skills from prompts.chat registry using MCP tools like search_skills and get_skill. Activates for finding skills, browsing catalogs, or extending Claude.
Creates, edits, and optimizes skills for Claude Code, including drafting, evaluating with test prompts, iterating on performance, and improving skill descriptions for better triggering accuracy.