"Explore code relationships, search, flows, communities,
Scanned 9/8/2026
Install to Claude Code
npx -y skills add VectorSpaceLab/AREX-Skill --skill graph-exploration --agent claude-codeInstalls into .claude/skills of the current project.
Are you the author of Graph Exploration?
Add the live security badge to your README — it updates automatically with every re-scan.
[](https://www.skillsdirectory.com/skills/vectorspacelab-graph-exploration)More formats (shields.io, HTML) on the badges page.
---
name: graph-exploration
description: "Explore code relationships, search, flows, communities,
architecture, and refactor previews with code-review-graph."
disable-model-invocation: true
metadata:
disco-role: operating
license: MIT
---
# Graph Exploration
Use this sub-skill when the task is to understand code structure after a graph exists: find callers or callees, search by keyword or embedding, inspect flows or communities, identify large functions, or preview safe refactors.
## Start here
1. Ensure the graph exists and is current.
2. Ask the smallest useful query first.
3. Escalate from search to graph traversal to flow/community/refactor analysis only as needed.
Read [references/search-and-analysis.md](references/search-and-analysis.md) for search, traversal, flow, community, architecture, and refactor workflows. Read [references/architecture-and-schema.md](references/architecture-and-schema.md) for graph-node and edge semantics, flow/community storage, and supported query shapes. Read [references/troubleshooting.md](references/troubleshooting.md) when search falls back, flows miss entry points, communities are sparse, or a refactor preview looks incomplete.
## Route by task
| User task | Do this |
| --- | --- |
| "Who calls this function?" | `query_graph_tool(pattern="callers_of", target=...)` |
| "What does this function call?" | `query_graph_tool(pattern="callees_of", target=...)` |
| "Find tests for this symbol" | `query_graph_tool(pattern="tests_for", target=...)` |
| "Search for login/auth code" | `semantic_search_nodes_tool` or `query_graph_tool` depending on whether the answer is semantic or structural. |
| "Show execution paths" | `list_flows_tool`, then `get_flow_tool` or `get_affected_flows_tool`. |
| "Show architecture" | `get_architecture_overview_tool` and, if needed, `list_communities_tool`. |
| "Find large functions" | `find_large_functions_tool`. |
| "Preview a rename" | `refactor_tool(mode="rename")`. |
| "Find dead code" | `refactor_tool(mode="dead_code")`. |
## Use search vs graph traversal correctly
- Use `semantic_search_nodes` for natural-language or fuzzy discovery.
- Use `query_graph_tool` when the question is about exact structural relationships.
- Use `traverse_graph_tool` when you already have a promising node and want to walk outward.
- Use `get_minimal_context_tool` first if you are unsure where to start.
## Flows and communities
Flows represent execution paths, usually rooted in entry points or decorators. Communities represent structural clusters and help with architecture-level reasoning. A good exploration answer usually explains both the local symbol and the surrounding flow/community context when relevant.
## Refactor previews
The safe refactor path is preview first, apply second:
1. `refactor_tool(mode="rename", ...)` or `refactor_tool(mode="dead_code", ...)`
2. Inspect the returned edits and affected locations.
3. Only then run `apply_refactor_tool` if the change is intentional.
## Boundaries
- For install/setup, use `install-and-setup`.
- For diff review and PR comments, use `review-changes`.
- For embeddings, custom languages, registry/daemon, wiki, and eval workflows, use `integrations-and-extensions`.
## Verification anchors
Native tests that ground this route include search tests, flow tests, community tests, integration tests, refactor tests, visualization tests, and status/regression tests. These are validation anchors only; do not rely on them as runtime instructions.Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.
No comments yet. Be the first to comment!