kit-cli-navigator
Run focused `kit` CLI commands (file-tree, symbols, search, semantic discovery, dependency analysis, and exports) to build high-signal repository context fast. Use when you need to understand an unfamiliar codebase, prep context for Claude, plan refactors, audit dependencies, or answer “where is X implemented?” without manually grepping.
85
82%
Does it follow best practices?
Impact
81%
3.85xAverage score across 3 eval scenarios
Passed
No known issues
Quality
Discovery
100%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 an excellent skill description that clearly identifies the specific tool (kit CLI), enumerates concrete capabilities, and provides an explicit 'Use when' clause with natural trigger scenarios. It uses third person voice throughout and covers both what the skill does and when to use it with strong, distinctive language.
| Dimension | Reasoning | Score |
|---|---|---|
Specificity | Lists multiple specific concrete actions: file-tree, symbols, search, semantic discovery, dependency analysis, and exports. Also specifies the tool ('kit' CLI) and the purpose ('build high-signal repository context fast'). | 3 / 3 |
Completeness | Clearly answers both 'what' (run focused kit CLI commands for file-tree, symbols, search, etc.) and 'when' (explicit 'Use when' clause covering understanding codebases, prepping context, planning refactors, auditing dependencies, and locating implementations). | 3 / 3 |
Trigger Term Quality | Includes strong natural trigger terms users would say: 'understand an unfamiliar codebase', 'plan refactors', 'audit dependencies', 'where is X implemented', 'grepping', 'search', 'symbols', 'dependency analysis'. These cover a wide range of natural user queries about code exploration. | 3 / 3 |
Distinctiveness Conflict Risk | Clearly scoped to the 'kit' CLI tool for repository context building, which is a distinct niche. The specific tool name and the enumerated commands (file-tree, symbols, semantic discovery) make it unlikely to conflict with generic code search or analysis skills. | 3 / 3 |
Total | 12 / 12 Passed |
Implementation
64%Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.
This is a solid, actionable skill with concrete CLI commands and well-structured examples that would genuinely help Claude navigate codebases using kit. Its main weaknesses are moderate verbosity from repeated subagent orchestration instructions woven into every step, and the lack of inline validation/error-recovery checkpoints within the core workflow. The progressive disclosure to reference.md is appropriate but the main document could be tighter.
Suggestions
Integrate validation checkpoints directly into the core workflow steps (e.g., 'If kit symbols returns empty, verify the path and language support before proceeding') rather than relegating all error handling to a separate pitfalls section.
Reduce repetition of the subagent pattern—state it once at the top of the workflow (as partially done) and remove the 'Subagent:' prefix from every individual step to save ~20% of tokens.
Move some of the per-command flag detail (e.g., --chunk-by, --format options, --cycles) into reference.md and keep the core workflow focused on the sequence and decision points.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The skill is reasonably efficient but includes some unnecessary repetition (e.g., repeatedly mentioning 'Subagent:' prefixes and the subagent orchestration pattern throughout every step). The 'When to activate' section and some best practices restate things Claude could infer. However, the command examples themselves are lean and useful. | 2 / 3 |
Actionability | Nearly every step includes concrete, executable CLI commands with realistic flags and arguments. The three worked examples provide end-to-end sequences with specific commands, output paths, and clear goals. Commands are copy-paste ready with real flag combinations. | 3 / 3 |
Workflow Clarity | The 7-step core workflow is clearly sequenced and logically ordered from scoping through packaging. However, there are no explicit validation checkpoints or error-recovery feedback loops within the workflow steps themselves—pitfalls are listed separately but not integrated as 'if X fails, do Y' within the numbered steps. For a skill that generates artifacts and runs multiple CLI commands, inline validation would strengthen it. | 2 / 3 |
Progressive Disclosure | The skill references 'reference.md' for flag-level detail, which is good progressive disclosure. However, no bundle files were provided, so we can't verify the reference exists. The main body is quite long (~150+ lines of detailed workflow) and some of the per-command detail could be offloaded to the reference sheet. The examples section adds value inline but the overall document is heavier than ideal for a top-level SKILL.md. | 2 / 3 |
Total | 9 / 12 Passed |
Validation
100%Checks the skill against the spec for correct structure and formatting. All validation checks must pass before discovery and implementation can be scored.
Validation — 11 / 11 Passed
Validation for skill structure
No warnings or errors.
- Repository
- cased/claude-code-plugins
- Commit
a026516
- 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.