Content
65%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.
The body is highly actionable with concrete, executable Membrane CLI commands and a well-instrumented connection state machine, but it is padded with explanatory prose, has inconsistent step numbering, and inlines reference material that a >50-line skill would benefit from splitting into separate files.
Suggestions
Trim explanatory Membrane-behavior prose (e.g. 'so you can focus on the integration logic rather than auth plumbing', the URL-normalization paragraph) to assume Claude's competence and reduce tokens.
Fix the step numbering: add an explicit Step 1 / Step 2 structure so the 'skip to Step 2' and '1b' references resolve, and add a verification checkpoint after running actions that modify Mati records.
Move the proxy flags table and the detailed clientAction field schema into a separate reference file (e.g. reference.md) referenced from the main body, so SKILL.md stays a lean overview.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is mostly efficient with concrete commands, but includes unnecessary prose such as 'so you can focus on the integration logic rather than auth plumbing' and explanatory Membrane-behavior paragraphs ('The URL is normalized to a domain and matched against known apps...') that could be trimmed, fitting the 3-anchor. | 3 / 5 |
Actionability | The body provides fully executable, copy-paste-ready commands throughout (install, login, connection ensure/get with --wait, action list/run, request proxy) complete with flags, --json output, and a flags table, matching the 5-anchor for fully executable guidance covering common cases. | 5 / 5 |
Workflow Clarity | The connection flow has good state-based checkpoints (READY/BUILDING/CLIENT_ACTION_REQUIRED/CONFIGURATION_ERROR with polling feedback), but the step numbering is inconsistent (a '1b' with no '1a', dangling 'skip to Step 2' references with no Step 2 header) and the action-running workflow lacks post-execution verification for what may be destructive record changes, triggering the destructive-operation cap at 3. | 3 / 5 |
Progressive Disclosure | The content is sectioned with clear headers and avoids nested references, but at ~140 lines with detailed reference material (the proxy flags table, the full clientAction sub-field schema) inlined into a single SKILL.md with no separate reference files, it fits the 3-anchor where content that should be split remains inline. | 3 / 5 |
Total | 14 / 20 Passed |