the-explore-codebase
Explore an unfamiliar area of the Stave codebase quickly. Use when the request asks for workspace structure, architecture overview, relevant files, entrypoints, "where should I look first", "관련 코드 확인", "구조 파악", or broad code search before implementation.
89
84%
Does it follow best practices?
Impact
99%
1.05xAverage score across 3 eval scenarios
Passed
No known issues
Quality
Discovery
89%Based on the skill's description, can an agent find and select it at the right time? Clear, specific descriptions lead to better discovery.
This is a strong skill description with excellent trigger term coverage (including bilingual terms) and a clear 'Use when' clause that makes it easy for Claude to select appropriately. The main weakness is that the 'what' portion could be more specific about the concrete actions the skill performs beyond just 'explore.' Adding explicit actions like 'maps module dependencies, identifies key entrypoints, summarizes directory structure' would strengthen it.
Suggestions
Add more concrete action verbs describing what the skill does, e.g., 'Maps module dependencies, identifies key entrypoints, summarizes directory structures, and locates relevant source files in the Stave codebase.'
| Dimension | Reasoning | Score |
|---|---|---|
Specificity | The description names the domain (Stave codebase exploration) and mentions some actions like 'workspace structure, architecture overview, relevant files, entrypoints,' but doesn't list multiple concrete actions the skill performs (e.g., generates dependency graphs, lists module hierarchies, traces call paths). | 2 / 3 |
Completeness | Clearly answers both 'what' (explore unfamiliar areas of the Stave codebase quickly) and 'when' with an explicit 'Use when...' clause listing multiple specific trigger scenarios. | 3 / 3 |
Trigger Term Quality | Excellent coverage of natural trigger terms including 'workspace structure', 'architecture overview', 'relevant files', 'entrypoints', 'where should I look first', Korean equivalents '관련 코드 확인' and '구조 파악', and 'broad code search before implementation'. These are phrases users would naturally say. | 3 / 3 |
Distinctiveness Conflict Risk | Highly distinctive — scoped specifically to the Stave codebase, focused on exploration/orientation rather than implementation, and includes unique triggers like 'where should I look first' and Korean phrases that clearly distinguish it from implementation or editing skills. | 3 / 3 |
Total | 11 / 12 Passed |
Implementation
79%Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.
This is a well-crafted exploration skill that is concise, actionable, and provides specific file paths and tool guidance for navigating the Stave codebase. Its main weakness is the workflow section, which lists steps but lacks explicit validation or decision points for confirming the right subsystem has been identified. The progressive disclosure is adequate with references to architecture docs and retrieved context, though the referenced files aren't bundled.
Suggestions
Add a brief validation checkpoint in the workflow, e.g., 'After step 2, confirm the subsystem matches the user's question before diving into files — if uncertain, check a second subsystem's entrypoints.'
Consider adding a decision tree or brief conditional guidance for when the retrieved context (`stave:repo-map`) is sufficient vs. when a full exploration agent is needed.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The content is lean and efficient. It assumes Claude knows how to use tools like `rg`, `Glob`, and `Agent`, and doesn't waste tokens explaining what they are. Every section earns its place — hotspots provide specific file paths with brief justifications, and the avoid section is terse but clear. | 3 / 3 |
Actionability | Provides specific file paths (e.g., `src/store/app.store.ts`, `electron/main/ipc/schemas.ts`), concrete tool preferences (`Agent` with `subagent_type: "Explore"`), and explicit guidance on when to use which approach. The hotspots section gives precise, actionable starting points with context on why each file matters. | 3 / 3 |
Workflow Clarity | The workflow steps are clearly sequenced (read docs → choose subsystem → identify files → explore → read slices), but there are no validation checkpoints or feedback loops. For an exploration task this is less critical than for destructive operations, but the skill doesn't specify how to verify you've found the right subsystem or when to stop exploring. | 2 / 3 |
Progressive Disclosure | The skill references external docs (`docs/architecture/index.md`, `docs/architecture/entrypoints.md`, `AGENTS.md`) and retrieved context sources, which is good navigation. However, there are no bundle files provided to support these references, and the hotspots section is inline content that could potentially be split out if it grows. The structure is reasonable but the references are one-directional without confirmation they exist. | 2 / 3 |
Total | 10 / 12 Passed |
Validation
90%Checks the skill against the spec for correct structure and formatting. All validation checks must pass before discovery and implementation can be scored.
Validation — 10 / 11 Passed
Validation for skill structure
| Criteria | Description | Result |
|---|---|---|
frontmatter_unknown_keys | Unknown frontmatter key(s) found; consider removing or moving to metadata | Warning |
Total | 10 / 11 Passed | |
- Repository
- sendbird/stave
- Commit
e275ad5
- Reviewed
Table of Contents
Is this your skill?
If you maintain this skill, you can claim it as your own. Once claimed, you can manage eval scenarios, bundle related skills, attach documentation or rules, and ensure cross-agent compatibility.