Skip to content

datit309/supergraph / plugins/supergraph/skills/explore

explore

Verified

Deep codebase exploration, architectural research, and solution discovery. Use when investigating problems, understanding complex flows, or evaluating technical approaches without jumping straight into code mutation or TDD.

FileKindSize
SKILL.mdskill md2.8 KB
---
name: explore
description: Deep codebase exploration, architectural research, and solution discovery. Use when investigating problems, understanding complex flows, or evaluating technical approaches without jumping straight into code mutation or TDD.
mcp: codebase-memory-mcp
---

# /supergraph:explore

Investigate first. Understand deeply. No unwanted code modification.

Announce: "🧭 /supergraph:explore — investigating codebase and tracing flows..."

## When to Use

- Understanding an existing feature, module, or architecture
- Tracing end-to-end request/event lifecycles (UI → BLoC/Service → API → DB)
- Investigating bug symptoms before declaring a fix approach
- Evaluating libraries, architectural options, or technical trade-offs
- Onboarding to a new workspace or subsystem

## Workflow

### 1. Define Research Scope
- **Subject**: What system, flow, or question is being investigated?
- **Boundaries**: Known entry points, affected modules, or target concepts.
- **Target Output**: Architecture overview, flow diagram, root cause analysis, or comparison matrix.

### 2. Semantic & Graph Discovery
- **High-level map**: Use `get_architecture(project=CBM_PROJECT)` or `/zoom-out` to orient.
- **Symbol / Route discovery**: Use Serena `find_symbol`, `get_symbols_overview` or Codebase Memory `search_graph`.
- **Trace connections**: Use `trace_path` or Serena `find_referencing_symbols` / `find_implementations` to identify callers, callees, and data flow.

### 3. Targeted Source Deep-Dive
- Inspect the actual code files for key components (controller, service, repository, UI state).
- Read method implementations directly when business logic, validations, or transformations matter.
- Target reading precisely: understanding requires seeing concrete implementations, not just graph nodes.

### 4. Trace the Flow End-to-End
Trace the execution path:
1. **Entry Point**: HTTP route handler, UI widget event, CLI command, or queue consumer.
2. **Middleware / Guards**: Authentication, validation, rate limiting, interceptors.
3. **Domain Logic / Orchestration**: Service layer, state management, entity calculations.
4. **Data / External I/O**: Database queries, external API calls, cache lookups.
5. **Output / Side Effects**: Response serialization, state updates, event emissions.

### 5. Synthesize & Report Findings
Present findings in clear, structured Markdown:
- **Executive Summary**: 2-3 sentences explaining how it works or what the finding is.
- **Flow Diagram**: ASCII or Mermaid sequence/flowchart.
- **Key Files & Responsibilities**: Table of files with their exact role.
- **Critical Invariants & Edge Cases**: Hidden assumptions, potential pitfalls, race conditions.
- **Recommendations**: If leading into implementation, outline next steps (`/plan`, `/analyze`, or `/prototype`).
add-product-e2e-evalpaperclipai/paperclipAdd or extend a Paperclip full-stack runner E2E workflow, fixture, matcher, or report evidence path for local or Daytona execution.Stars: 115kUpdated Claude CodeCodexVerifiedInstall
add-runner-evalpaperclipai/paperclipAdd or extend a Paperclip Runner protocol evaluation definition, roster, assertion, or report fixture with provenance and narrow validation.Stars: 115kUpdated Claude CodeCodexVerifiedInstall
diagnose-why-work-stoppedpaperclipai/paperclipCompatibility entry point for failed, stalled, looping, or over-recovered Paperclip work. Use fix-failed-paperclip-run for the full diagnosis, regression, repair, real verification, and green-PR workflow; honor an explicitly requested diagnosis-only scope.Stars: 115kUpdated Claude CodeCodexVerifiedInstall
fix-failed-paperclip-runpaperclipai/paperclipDiagnose and fix failed, stuck, slow, or incorrectly blocked Paperclip tasks and runs. Use with a task/run URL or failure report to trace the root cause, prove a regression red then green, check second-order effects, verify the real workflow, and deliver a PR with passing checks.Stars: 115kUpdated Claude CodeCodexVerifiedInstall