Content
75%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 comprehensive, executable Membrane CLI commands and a well-sequenced connection workflow that includes genuine validation/feedback loops. Its main weaknesses are an unnecessary product-introduction paragraph, a broken "Official docs" line, and slightly inconsistent step numbering.
Suggestions
Remove or trim the introductory product-description paragraph and the broken "Official docs: https://www.e" line; assume Claude can infer the product's purpose.
Reconcile the step labels so referenced steps ("Step 2," "1b") have matching headers or are renamed to consistent steps.
Consider moving the verbose connection-state reference (clientAction sub-fields) and the proxy flag table into a reference file if the body grows, to keep SKILL.md a lean overview.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | Most of the body is lean, executable CLI commands, but the opening "Time Tracker by eBillity is a time tracking and expense management software. It's used by businesses of all sizes to monitor employee work hours, track project costs, and generate invoices." explains the product at a level that is largely unnecessary, and the dangling "Official docs: https://www.e" line is broken padding. Mostly efficient with some trimming needed, matching anchor 3. | 3 / 5 |
Actionability | The body is packed with copy-paste-ready commands across the full workflow — install, login, connection ensure, --wait polling, action list/run with --input JSON, and membrane request with a full flag table. These are concrete, executable, and cover the common cases per anchor 5; placeholders like CONNECTION_ID are standard substitution, not pseudocode. | 5 / 5 |
Workflow Clarity | There is a clear sequenced workflow (install → authenticate → ensure connection → wait/poll for READY → search actions → run → proxy) with real validation checkpoints: polling --wait until state leaves BUILDING, branching on READY/CLIENT_ACTION_REQUIRED/CONFIGURATION_ERROR, and re-polling after the user acts. Not a 5 because the step numbering is inconsistent ("Step 2" and "1b" are referenced but not cleanly labeled), leaving minor sequence gaps. | 4 / 5 |
Progressive Disclosure | No bundle files exist and none are needed; the content is organized into clear, well-labeled sections (Overview, Install, Authentication, Connecting, Searching, Running actions, Proxy, Best practices). It is well-structured and navigable inline, fitting anchor 4; not a 5 because the detailed connection-state handling and proxy flag table could be separated, but keeping them inline is defensible. | 4 / 5 |
Total | 16 / 20 Passed |