Content
38%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 CLI workflow sections are genuinely executable and well-specified, but they are buried under a massive, meaningless entity list that destroys token efficiency and structure, and the step numbering is broken. Removing the list and fixing the cross-references would transform this from a failing skill into a solid one.
Suggestions
Delete the ~390-line 'BusinessLogic Overview' entity bullet list entirely; it is padding with no instructional value (and contains duplicates like Comment, Report, Dashboard, Success).
Fix the workflow cross-references: define the steps the text points to ("skip to Step 2" has no Step 2; "1b. Wait for the connection" has no 1a/Step 1) or replace them with descriptive section links.
Add one worked example that runs a listed action end-to-end, e.g. membrane action run execute-business-logic --connectionId=... --input '{...}' --json, so users see how an action key maps to an actionId.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body contains a ~390-line bullet list of irrelevant nouns ("Dream", "Nightmare", "Love", "One" through "Trillion", "Universe", "Everything", "Nothing") that carries no instructional value — severely padded, worse than the anchor-1 example of over-explaining known concepts. Every remaining section would need to be near-perfect to escape the bottom anchor, and the padding dominates the file. | 1 / 5 |
Actionability | The operational sections give copy-paste-ready commands: "npm install -g @membranehq/cli@latest", "membrane connection ensure \"https://www.businesslogic.online/\" --json", "membrane action run <actionId> --connectionId=CONNECTION_ID --input '{\"key\": \"value\"}' --json", plus a complete proxy flag table. Minor gaps keep it below 5: "Use action names and parameters as needed" is vague, and no worked example invokes one of the two listed actions with its real key. | 4 / 5 |
Workflow Clarity | A real sequence exists (install → login → connection ensure → poll state → search actions → run) with state checkpoints (READY/BUILDING/CLIENT_ACTION_REQUIRED and CONFIGURATION_ERROR handling), but "skip to **Step 2**" points to a step that is never defined, and a "1b" heading appears with no "1a" or "Step 1" counterpart. The broken cross-references and implicit search-to-run handoff match 'sequence present but checkpoints missing or implicit'. | 3 / 5 |
Progressive Disclosure | No bundle files exist, yet ~390 lines of entity-list content that clearly belongs in a separate file (or nowhere) are inlined into SKILL.md, matching 'content that clearly belongs in separate files is inlined'. The operational sections themselves have reasonable headers, placing this above the structureless anchor 1 but below anchor 3, whose reference/organization expectations are irrelevant given there is nothing to reference and the inline bulk is the dominant structural problem. | 2 / 5 |
Total | 10 / 20 Passed |