Conciseness

The bulk is lean, copy-paste commands and tables, but there is real padding: the intro paragraph 'Carbone is a reporting engine... Developers and businesses use it to automate report creation', the skeletal Overview bullets (Template/Report/Report Group), and the vacuous line 'When to use which actions: Use action names and parameters as needed.' This fits anchor 3 (mostly efficient, some unnecessary explanation) rather than 4, since the Overview section is more than a minor instance of over-explanation.

Actionability

Fully executable, copy-paste-ready commands throughout: 'npm install -g @membranehq/cli@latest', 'membrane login --tenant --clientName=<agentType>', 'membrane connection ensure "https://carbone.io/" --json', 'membrane action run <actionId> --connectionId=CONNECTION_ID --input {\'key\': \'value\'}', plus the proxy request syntax with a flags table and a Popular actions table with concrete action keys. This matches anchor 5; not 4 because the examples cover the common cases (install, auth, connect, search, run, proxy) with exact, runnable syntax.

Workflow Clarity

The sequence install → login → connection ensure → poll state → search actions → run → proxy fallback is clear, with explicit checkpoints and a genuine feedback loop ('poll again with membrane connection get <id> --json to check if the state moved to READY') and per-state error handling (CLIENT_ACTION_REQUIRED, CONFIGURATION_ERROR, SETUP_FAILED). It misses 5 because steps are inconsistently labeled — the text says 'skip to **Step 2**' and '#### 1b' but no step is actually numbered/labeled, leaving a dangling reference.

Progressive Disclosure

The single-file skill (no bundle files exist) is well-sectioned with clear headers (Install, Authentication, Connecting, Searching, Popular actions, Running, Proxy, Best practices), and its size (~145 lines) is reasonable for inline CLI usage. It fits anchor 4 (good structure, minor organization gaps) rather than 5 — the file exceeds the ~50-line simple-skill case, and the popular-actions and proxy-flag tables plus the step-labeling inconsistency leave minor organization gaps.

Specificity

"Manage data, records, and automate workflows" names the Carbone domain but the actions are entirely generic — no concrete capability like rendering documents from templates, uploading templates, or generating reports appears. It matches anchor 2 ('Names the domain but actions are minimal or generic') rather than 3, because not even 1-2 concrete actions are listed.

Completeness

The 'when' is explicitly present ("Use when the user wants to interact with Carbone data"), but the 'what' ("Carbone integration. Manage data, records, and automate workflows") is vague — it never says the skill generates documents/spreadsheets from templates and JSON data. It sits between anchors 3 and 4; the vague 'what' undercuts it more than the generic 'when' helps, so it cannot reach 4 ('both what and when; when could be more explicit' assumes a real what).

Trigger Term Quality

It has the distinctive natural keyword "Carbone" plus "data", "records", "workflows", but misses the natural phrases users of this service would say: "reports", "documents", "templates", "render", "generate". This fits anchor 3 (some relevant keywords, missing common variations) rather than 4, since coverage of natural terms is thin.

Distinctiveness Conflict Risk

"Carbone integration" and "interact with Carbone data" establish a clear niche with a distinct app-name trigger, giving minor conflict risk. Not 5 because the generic phrases "Manage data, records, and automate workflows" overlap with data-management and workflow-automation skills; not 3 because the Carbone name is far more distinctive than the anchor-3 example 'Works with document files'.