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.
A dense, highly actionable reference that assumes Claude's competence: complete titled code examples for every function variant, platform limits and gotchas stated numerically, and deliberate first-example design. The main weaknesses are the appended house-rules section that duplicates ~60% of its rules from the body, the absence of an explicit ordered build-test-deploy workflow, and a monolithic single-file structure with no reference split. All issues are polish-level; the skill would score near the top of the range after trimming the duplicate section and extracting reference detail.
Suggestions
Trim or deduplicate the appended 'Netlify house rules' section: rules 5–13 restate warnings already present in the body (geo mocking, memory/vcpu, region, included_files, the 4 KB env cap, cron UTC conversion, header scope); keeping each fact in one place would cut roughly 40 lines without losing information.
Add a short ordered workflow near the top or in 'Local testing & deploy' — write the function → run `netlify dev` and exercise it locally (including `--geo=mock`) → deploy via Git or CLI → verify in the Functions tab / streamed logs — so the end-to-end sequence is explicit rather than implied by section order.
Consider moving the platform-event handler tables, region/memory reference detail, and the legacy Lambda-compat section into a references/ file linked from SKILL.md, keeping the main file as a leaner overview.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The main body is dense and warning-driven with essentially zero explanation of concepts Claude already knows (no "what is serverless" padding; every line carries Netlify-specific facts like "60s execution, 20 MB response" and the ~4 KB env-var cap). The deduction is the appended ~70-line house-rules section where rules 5–13 restate warnings already given above (geo mocking, memory/vcpu, region, included_files, the 4 KB cap, cron UTC conversion, header scope), which is more than a minor trim candidate. This sits above the 3 anchor (mostly efficient) because the duplication reinforces facts rather than padding with known-concept explanation, but below the 5 anchor because those ~70 lines don't each earn their place. | 4 / 5 |
Actionability | Nearly every function variant ships a complete, copy-paste-ready example with its file path in the code fence title ("netlify/functions/hello.mts", "netlify/functions/daily-digest.mts"), plus executable commands like `netlify dev --geo=mock --country=DE` and `netlify functions:invoke <name>` and runnable `netlify.toml` snippets. Common cases (minimal function, path routing, scheduled, background, identity check, legacy migration) are each covered end to end. The only placeholder is the streaming example's `messages: [/* ... */]`, which is inherent to the example; this matches the 5 anchor rather than the 4 anchor. | 5 / 5 |
Workflow Clarity | The content is well-sequenced for its reference nature: modern default handler first, file locations, then a deliberately minimal first example ("No `config` export. Serves at `/.netlify/functions/hello`") before advanced config, with a dedicated "Local testing & deploy" section covering framework dev servers, `netlify dev`, and deploy via Git/CLI. Redeploy checkpoints are stated for env var, region, and runtime changes, and a diagnostic loop exists for background-flag version skew ("check and upgrade the local tooling"). It falls short of the 5 anchor because there is no explicit ordered write → test → deploy → verify-in-logs workflow with validation feedback loops; it stays above the 3 anchor because sequence and most checkpoints are present and no destructive/batch cap applies. | 4 / 5 |
Progressive Disclosure | No bundle files exist and the body references none (only two external doc URLs), so there are zero nested-reference levels and nothing is buried; navigation is easy via ~15 clearly headed sections ordered from basics to legacy. This matches the 4 anchor (good structure, content mostly appropriately placed, minor organization gaps): the gap is that the single monolithic ~320-line file inlines material that could be split out (platform-event handler tables, region/memory reference detail, the legacy Lambda-compat section). It is above the 3 anchor — whose example is unstructured bulk with unclear references — because sectioning is clean and scannable, but below the 5 anchor, whose profile is a concise overview with well-signaled one-level references. | 4 / 5 |
Total | 17 / 20 Passed |