Content
71%Weight 40%Scale 1-5Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.
A well-organized, token-efficient instruction body with a clear session workflow and a sensible diagnostic path for missing tools. Its main gap is actionability: the build/run/test tools that the whole skill exists to orchestrate are referenced only obliquely, with no example invocations, leaving Claude to guess tool names from the MCP tool list.
Suggestions
Name the concrete tools for the core workflows — especially the 'combined build-and-run tool' and the build/test/log/debug/UI-automation tools mentioned in the capabilities list — so the guidance is executable without guessing from the tool registry.
Add one minimal example tool invocation (e.g., a `session_show_defaults` call and a build-and-run call with typical parameters) to anchor the workflow steps in concrete usage.
De-duplicate the tool-availability explanation (the intro paragraph vs. Step 2) into a single location to tighten conciseness.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is lean and directive ("Use MobileBuildMCP tools instead of raw xcodebuild, xcrun, or simctl"; terse step bullets) without explaining concepts Claude already knows. Minor trimming opportunities exist: tool availability is explained twice (the 'Only simulator workflow tools are enabled by default' paragraph and Step 2), and a long external GitHub URL consumes tokens. This fits anchor 4 (efficient; minor over-explanation) rather than anchor 5. | 4 / 5 |
Actionability | Some concrete guidance exists — named tools `session_show_defaults` and `discover_projs`, and the config path `.mobilebuildmcp/config.yaml` — but the core workflow tools for build/run/test are never named ("prefer the combined build-and-run tool" gives no tool name), and there are no example tool invocations or parameter patterns. Key details are missing, matching anchor 3 rather than anchor 4's 'minor gaps'. | 3 / 5 |
Workflow Clarity | Steps 1–3 form a clear sequence (establish context → check workflow availability → report context) with ordering rules ("Do not run discovery speculatively") and a diagnostic loop (tool missing → check enabled workflows → update config → reload session). It falls short of anchor 5 because there are no explicit verification checkpoints before acting (e.g., confirming defaults are correct before building) and failure recovery is only summarized in Step 3. | 4 / 5 |
Progressive Disclosure | The skill is a single file under 50 lines with no bundle directories (references/, scripts/, assets/ are absent) and no need for external references; the capabilities list is compact enough to belong inline. Per the simple-skill exception, well-organized sections (intro, capabilities, Steps 1–3) warrant a 5. The external configuration URL is a documentation pointer, not a nested local reference. | 5 / 5 |
Total | 16 / 20 Passed |