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.
This is a well-structured, command-driven integration skill with a strong connection state machine and concrete CLI examples throughout. Its main weaknesses are a minor invocation inconsistency, a lack of validation guidance on the destructive/proxy side, and a low-value entity list that adds tokens without explanation.
Suggestions
Standardize the CLI invocation (pick either 'membrane' or 'npx @membranehq/cli@latest' and use it consistently, noting the alias once).
Add a validation/verification checkpoint before destructive proxy calls (e.g. 'For -X DELETE, confirm the target id and dry-run or list the resource first').
Either annotate the Mav Overview entity list with one-line descriptions or move it to a reference file so the bullets earn their tokens.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is mostly efficient with concrete commands and a focused flags table, though the unexplained entity bullet list and the product-intro paragraph add minor tokens that could be trimmed, fitting 'efficient; minor instances of over-explanation'. | 4 / 5 |
Actionability | It gives copy-paste-ready commands for install, login, connection ensure/get, action list/run, and request with flag tables, but the invocation form is inconsistent ('membrane' vs 'npx @membranehq/cli@latest action list'), a minor gap keeping it just below fully executable. | 4 / 5 |
Workflow Clarity | The connection flow is a clear state machine (READY/BUILDING/CLIENT_ACTION_REQUIRED/CONFIGURATION_ERROR) with polling and a re-auth feedback loop, but the action-running and proxy (DELETE) side lacks explicit validation checkpoints, leaving minor validation gaps. | 4 / 5 |
Progressive Disclosure | With no bundle files present, the single SKILL.md is well organized into clear sections (Overview, Working with Mav, Install, Auth, Connecting, Searching, Proxy, Best practices) and content is appropriately inline, with only minor organization gaps like the unexplained entity list. | 4 / 5 |
Total | 16 / 20 Passed |