From pactkit
Traces code execution flows via static analysis, generates call graphs and Mermaid sequence diagrams. Use to map existing logic before changes or confirm call sites.
How this skill is triggered — by the user, by Claude, or both
Slash command
/pactkit:pactkit-traceThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
Deep code analysis and execution path tracing via static analysis.
Deep code analysis and execution path tracing via static analysis.
Grep to locate entry points (API route, CLI arg, Event handler).visualize --mode call --entry <function_name> to obtain call chains.docs/architecture/graphs/call_graph.mmd to see all reachable functions.dict -> UserObj -> JSON).Output a Mermaid Sequence Diagram to visualize the flow.
If detect_topology(root) returns topologies beyond PDCA/Service:
Frontend API Topology (if api_call detected):
api_convention_summary(root) to get path prefixes, fetch function names, total call count./api/v1/) and fetch wrapper (e.g., apiFetch).[dynamic] markers) that may need special handling.Agent Topology (if agent detected):
StateGraph (stdlib ast), YAML agent definitions, MCP server configs.orchestrates edges in the report.npx claudepluginhub pactkit/claude-code-plugin --plugin pactkitTraces code execution paths, data flows, and call graphs via /sourceatlas:flow queries. Reveals boundaries, entry points, cycles for understanding features and implementations.
Traces feature, request, or data flows through codebase from entry points (API routes, UI handlers, jobs) to storage, documenting layers, transformations, side effects, and dependencies via graphs or grep.
Traces execution paths, maps dependencies, follows data flows, and explores unfamiliar code systematically from entry points to build incremental understanding.