Content
82%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 a strong, execution-first skill: every step has a copy-paste command, and connection setup includes an explicit state machine with recovery paths. The main flaws are broken internal step references and the absence of any validation guidance for action-run results.
Suggestions
Fix the dangling navigation: "skip to **Step 2**" and the "#### 1b." heading reference steps that are never labeled — either number the top-level sections (Step 1: connect, Step 2: search/run actions) or drop the references.
Add a short validation step after `membrane action run`, e.g. check the response for an `error` field before reporting the `output` field as the result.
Trim the marketing-style padding ("It's used by marketers, sales teams, and data analysts...", "so you can focus on the integration logic rather than auth plumbing") to pure operational guidance.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is dominated by executable commands and compact tables with almost no concept explanation; the two-sentence platform intro and phrases like "so you can focus on the integration logic rather than auth plumbing" are minor padding that could be trimmed. Fits anchor 4 ('efficient; minor instances of over-explanation') rather than 5, where every token earns its place. | 4 / 5 |
Actionability | Fully executable, copy-paste-ready commands cover every phase: install (`npm install -g @membranehq/cli@latest`), auth including the headless completion flow, connection setup with polling (`--wait`), action discovery (`membrane action list --intent ...`), action execution with `--input` JSON, and proxy requests with a flags table. The Popular Actions table with action keys and descriptions covers the common cases, matching anchor 5. | 5 / 5 |
Workflow Clarity | A clear install→authenticate→connect→search→run sequence exists with explicit state checkpoints (READY / BUILDING / CLIENT_ACTION_REQUIRED / CONFIGURATION_ERROR) and feedback loops (re-poll after user action, check the `error` field on failure). Not 5 because the step labeling is inconsistent — the body says "skip to **Step 2**" and uses "#### 1b" but no Step 1 or Step 2 headings exist — and `action run` results lack any output/error validation guidance. | 4 / 5 |
Progressive Disclosure | No bundle files exist and the body references none, so there are no dangling or nested references; the single file is well-sectioned at a reasonable length. Fits anchor 4 ('good structure; most content is appropriately placed') rather than 5, since the clientAction state-machine detail and proxy options table are candidates for one-level-deep reference files as the skill grows. | 4 / 5 |
Total | 17 / 20 Passed |