Content
75%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 highly actionable with copy-paste Membrane CLI commands and a well-validated connection workflow, though it opens with unnecessary background on what Serverless is. Structure is good but the long inline resource list and an undefined 'Step 2' reference leave minor gaps.
Suggestions
Cut the opening paragraph explaining what Serverless/AWS Lambda is — Claude already knows this — and lead directly with the Membrane interaction model.
Resolve the dangling 'Step 2' reference and consistent step numbering (define 1a/Step 2 or relabel) so the connection workflow is unambiguous.
Move the 60-item resource type list into a separate references file and keep SKILL.md as an overview, or collapse it to the few types users most commonly act on.
Add a brief verification note for proxy requests (e.g. confirm the response status before treating a destructive DELETE/PUT as succeeded), since the proxy path supports arbitrary HTTP methods.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The opening paragraph ('Serverless is a development framework for building applications on serverless computing platforms like AWS Lambda...') explains a concept Claude already knows, and phrases like 'so you can focus on the integration logic rather than auth plumbing' pad the text, matching 'mostly efficient but includes some unnecessary explanation'; not a 2 because the bulk is lean executable commands. | 3 / 5 |
Actionability | Copy-paste ready commands cover the common cases — 'npm install -g @membranehq/cli@latest', 'membrane connection ensure "https://serverless.com" --json', 'membrane action run <actionId> --connectionId=CONNECTION_ID --input ...' — plus a flag table, fitting 'fully executable; copy-paste ready commands; specific examples cover the common cases'. | 5 / 5 |
Workflow Clarity | The connect-then-poll flow has explicit validation checkpoints ('If the returned connection has state: READY, skip to Step 2', poll until not BUILDING, state-driven next actions) and a feedback loop (poll again after the user authenticates), but the undefined 'Step 2' reference and inconsistent numbering ('1b' without a clear 1a/Step 2) leave minor gaps, matching anchor 4. | 4 / 5 |
Progressive Disclosure | Content is organized into clear, navigable sections (Authentication, Connecting, Searching, Running actions, Proxy, Best practices) with no nested references; however the ~60-item inline resource list ('Function, Version, Alias, ...') is a long inlined block that could live in a separate reference file, leaving it at 'good structure; minor organization gaps' rather than a 5. | 4 / 5 |
Total | 16 / 20 Passed |