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 is a solid, actionable integration guide: real commands, clear connection lifecycle with state-based checkpoints, and a proxy escape hatch. Its main weaknesses are a padded report catalog that belongs in a reference file, dangling 'Step 2' references with no matching headings, and the absence of a single worked end-to-end example.
Suggestions
Move the Ahrefs Overview report catalog (and optionally the popular-actions table) into a references/ file, keeping a short capability summary in SKILL.md with a clearly signaled link
Fix the dangling navigation: 'skip to **Step 2**' is referenced twice but no Step 2 heading exists, and '1b. Wait for the connection' implies a '1a' that is never labeled — number the workflow steps consistently
Add one worked end-to-end example, e.g. `membrane action run get-domain-rating --connectionId=... --input '{"target": "example.com"}' --json`, and clarify how the action table's keys map to <actionId>
Trim concept-explaining filler such as the opening paragraph on what Ahrefs is and the repeated 'Overview — provides a general overview' entries
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The command documentation is tight, but there is unnecessary explanation: the intro ("Ahrefs is a popular SEO tool suite used by marketers and SEO professionals...") and near-tautological catalog entries like "**Overview** — Provides a general overview of a website's SEO performance" repeated across three explorers. Mostly efficient but could be tightened, matching anchor 3 rather than 4 due to the volume of low-value report-catalog lines. | 3 / 5 |
Actionability | Concrete executable commands throughout — `npm install -g @membranehq/cli@latest`, `membrane connection ensure "https://ahrefs.com" --json`, `membrane action run <actionId> --input '{"key": "value"}'` — plus a proxy flag table. Mostly executable with minor gaps: no worked end-to-end example (e.g. running get-domain-rating against a real target) and the mapping from the action table's keys to <actionId> is implicit, so it does not reach anchor 5. | 4 / 5 |
Workflow Clarity | The sequence (install → login → connection ensure → poll state → search actions → run → proxy fallback) is clear, with explicit state checkpoints (READY, BUILDING, CLIENT_ACTION_REQUIRED, CONFIGURATION_ERROR) and polling guidance for error recovery. Not 5: "skip to **Step 2**" is referenced twice but no Step 2 heading exists and "1b" implies an unlabeled "1a", leaving minor incoherence in the sequence. | 4 / 5 |
Progressive Disclosure | A single-file skill with no bundle files; sections are well organized, but the ~35-line Ahrefs report catalog and the popular-actions table are reference-like content inlined in SKILL.md that would sit better in a separate file — matching anchor 3 (structure present, content that should be separate is inline). Not 2 because the body is navigable and well-sectioned; not 4 because at ~180 lines the report catalog and action reference clearly belong in a dedicated reference file. | 3 / 5 |
Total | 14 / 20 Passed |