Content
63%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 delivers strongly executable CLI guidance with a well-handled connection lifecycle and error states, and mostly respects token budget. Its main weaknesses are inlined bulk action-catalog content with no reference files, mild background padding, and broken step numbering (a 'Step 2' that doesn't exist, a '1b' with no '1a').
Suggestions
Fix the step references: label the connection-ensure section as Step 1 (rename '1b' accordingly) and add the missing 'Step 2' (searching/running actions) that READY-state readers are told to skip to.
Move the 20-row Popular actions table to a references/ file (e.g. references/actions.md) and keep only a handful of representative examples inline in SKILL.md.
Trim background padding — the 'Cisco Meraki provides cloud-managed IT solutions...' paragraph, the 'Use action names and parameters as needed.' filler line, and the 'Agent Types ... etc.' aside.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | Mostly efficient — commands and tables dominate — but there is unnecessary padding: the opening paragraph explaining what Cisco Meraki is ("Cisco Meraki provides cloud-managed IT solutions..."), the filler line "Use action names and parameters as needed.", and the loose "Agent Types ... etc." paragraph. Not score 4 because these sections explain context Claude already knows or add nothing actionable. | 3 / 5 |
Actionability | Commands are copy-paste ready throughout: `npm install -g @membranehq/cli@latest`, `membrane connection ensure "https://meraki.cisco.com/" --json`, `membrane action run <actionId> --connectionId=... --input '{...}'`, plus a flag table for proxy requests. Not score 5 because "skip to **Step 2**" points to a section that doesn't exist, and `npx @membranehq/cli` appears alongside `membrane` without reconciliation. | 4 / 5 |
Workflow Clarity | A clear sequence runs install → login (with headless fallback and `login complete <code>`) → connection ensure → poll with `--wait` → handle CLIENT_ACTION_REQUIRED/CONFIGURATION_ERROR states → search actions → run, and the connection state machine acts as a validation checkpoint. Not score 5 because step numbering is incoherent ("#### 1b" with no 1a, a referenced Step 2 that doesn't exist) and create/update actions get no verification guidance. | 4 / 5 |
Progressive Disclosure | Sections are well-labeled but there are no bundle files at all, and the 20-row Popular actions table is bulk catalog content inlined in SKILL.md that clearly belongs in a separate reference file. Not score 4 because content that should be split out is inline with no one-level-deep references; not score 2 because the inline content is still organized under clear headers. | 3 / 5 |
Total | 14 / 20 Passed |