Content
67%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, command-driven guide: every workflow step is executable, connection states and error recovery are handled explicitly, and the proxy fallback covers uncovered use cases. Its weaknesses are modest — a skippable what-is-Amplitude intro, a vestigial overview list, a dangling "Step 2" reference, and no worked example with real parameter values.
Suggestions
Delete the "Amplitude is a product analytics platform..." intro paragraph and the "Amplitude Overview" bullet list — Claude already knows what Amplitude is, and the "Use action names and parameters as needed" line adds nothing.
Fix the broken navigation: "skip to **Step 2**" and the "#### 1b." heading reference steps that are never defined — either number the workflow steps (1. Connect, 2. Search, 3. Run) or rewrite as plain cross-references to section names.
Add one concrete worked example, e.g. a full track-events run with a realistic JSON payload and expected output-field response, so the run pattern is demonstrated end-to-end rather than with '{"key": "value"}'.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The opening paragraph ("Amplitude is a product analytics platform that helps companies understand user behavior across their digital products. Product managers, marketers, and data scientists use it...") explains a concept Claude already knows, and the "Amplitude Overview" bullet list adds little, ending with the low-value "Use action names and parameters as needed." The rest is mostly commands and tables, so this is anchor 3 — mostly efficient with some unnecessary explanation to trim. | 3 / 5 |
Actionability | Nearly every step has a copy-paste-ready command (npm install, membrane login including headless completion, connection ensure, polling, action list/run, proxy flag table), which approaches anchor 5. It falls to 4 because no example uses concrete parameter values — the run example is the generic '--input "{\"key\": \"value\"}"' and no worked end-to-end action (e.g. a track-events payload) is shown. | 4 / 5 |
Workflow Clarity | The sequence install → login → connection ensure → poll state (with explicit READY / CLIENT_ACTION_REQUIRED / error-state handling and a re-poll loop) → search actions → run → proxy fallback is clear with most checkpoints present, matching anchor 4. It is not 5 because "skip to **Step 2**" references a step label that does not exist anywhere in the document, and there is no post-run verification guidance for action results. | 4 / 5 |
Progressive Disclosure | A single SKILL.md with no bundle files, organized under clear headers (Install, Authentication, Connecting, Searching, Popular actions, Running, Proxy, Best practices) that make navigation easy — good structure with minor organization gaps (anchor 4). It is not 5 because the 12-row popular-actions table and proxy flag table are reference material inlined in the overview file rather than split into a one-level-deep reference; it is not 3 because the sections are well-signaled and nothing is buried. | 4 / 5 |
Total | 15 / 20 Passed |