Build and reason about import graphs, call graphs, schema graphs, and service dependency graphs. Use when tracing paths, finding god nodes, surfacing cycles, or answering architecture questions about how parts connect.
npx claudepluginhub v1truv1us/ai-eng-system --plugin ai-eng-learningThis skill uses the workspace's default tool permissions.
Use structural graphs when the real question is "how do these parts connect?" Dependency graphs help answer path, coupling, boundary, and architectural gravity questions that plain search cannot answer reliably.
Creates isolated Git worktrees for feature branches with prioritized directory selection, gitignore safety checks, auto project setup for Node/Python/Rust/Go, and baseline verification.
Executes implementation plans in current session by dispatching fresh subagents per independent task, with two-stage reviews: spec compliance then code quality.
Dispatches parallel agents to independently tackle 2+ tasks like separate test failures or subsystems without shared state or dependencies.
Use structural graphs when the real question is "how do these parts connect?" Dependency graphs help answer path, coupling, boundary, and architectural gravity questions that plain search cannot answer reliably.
Nodes are files or packages. Edges are imports or requires.
Nodes are functions or methods. Edges are calls.
Nodes are tables, models, or fields. Edges are references and joins.
Nodes are services or jobs. Edges are network calls, events, or queue handoffs.
Before analyzing, decide:
Mixing edge semantics makes the graph misleading.
Start from the question:
Scope the graph to the relevant subsystem first.
Look for:
For architecture questions, path tracing is often more useful than global stats:
Summarize findings in actionable language:
| Signal | Typical Meaning |
|---|---|
| Very high out-degree | coordinator, god object, orchestration hotspot |
| Very high in-degree | shared dependency, utility choke point |
| High betweenness | bridge, hidden coupling, migration risk |
| Strong cycles | boundary leak, layered architecture failure |
| Surprise edge | cross-domain dependency worth review |
| Rationalization | Reality |
|---|---|
| "Search results are enough" | Search finds mentions, not actual structural paths. |
| "Only cycles matter" | High-centrality nodes can be just as risky as cycles. |
| "One huge graph is best" | Subsystem-scoped graphs are easier to trust and act on. |