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 content is strongly actionable and workflow-oriented: concrete tool sequences, parameter tables, an async polling pattern with validation checkpoints, and honest pitfalls throughout. Its weaknesses are redundancy (pitfalls duplicated between workflow sections and the 'Known Pitfalls' section, plus filler boilerplate in 'When to Use'/'Limitations') and the absence of any progressive-disclosure structure — everything is inlined in one long file. Splitting per-tool detail into reference files and de-duplicating the pitfalls would materially tighten it.
Suggestions
De-duplicate pitfalls: keep operational gotchas in the workflow sections OR in 'Known Pitfalls', not both, and remove the generic 'When to Use' boilerplate ("This skill is applicable to execute the workflow or actions described in the overview").
Split per-tool parameter detail (e.g. full AMPLITUDE_SEND_EVENTS and IDENTIFY payloads) into a references/ file, keeping SKILL.md as an overview with clearly signaled one-level-deep links.
Add one complete copy-paste-ready example tool call for the most common workflow (send events) including a realistic events array with user_id, event_type, and millisecond time.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is mostly efficient and free of explanations of concepts Claude already knows, but it is noticeably redundant: user-ID and timestamp pitfalls appear both in workflow 'Pitfalls' blocks and again verbatim-ish in 'Known Pitfalls', the Quick Reference table repeats tool/param listings, and the 'When to Use' section ("This skill is applicable to execute the workflow or actions described in the overview") is pure filler. It is not a 2 because the majority of sections are dense, useful operational content. | 3 / 5 |
Actionability | Each workflow gives a concrete numbered tool sequence, key parameters with types, a complete executable JSON example for IDENTIFY property operations, and step-by-step ID-resolution and polling patterns. It falls short of 5 only because there is no full example tool-invocation payload (e.g. a complete AMPLITUDE_SEND_EVENTS call) for the most common case. | 4 / 5 |
Workflow Clarity | Sequences are clear with explicit checkpoints ("Confirm connection status shows ACTIVE before running any workflows") and a genuine feedback loop for the async batch cohort operation ("Repeat step 2 until status is 'complete' or 'error'"), so the batch-operation cap at 3 does not apply. It stops short of 5 because error-recovery guidance beyond polling (e.g. what to do on 'error' status or failed event sends) is thin. | 4 / 5 |
Progressive Disclosure | Sections are well-organized with no buried or nested references, but this is a ~220-line single file with no bundle directories (references/, scripts/, assets/ do not exist), and per-tool parameter detail that belongs in a separate reference file is inlined in SKILL.md — matching 'some structure but content that should be separate is inline'. The under-50-line exception for a 5 does not apply. | 3 / 5 |
Total | 14 / 20 Passed |