Content
61%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 delivers concrete, executable CLI guidance with a well-validated connection workflow (state polling and client-action handling are genuinely strong). Weaknesses are a broken step numbering scheme (a dangling 'Step 2' reference), vague filler around the entity overview, no verification guidance for data-mutating actions, and a marketing paragraph that wastes tokens.
Suggestions
Fix the dangling "skip to **Step 2**" reference by numbering the workflow explicitly (1. Install, 2. Authenticate, 3. Connect, 4. Discover actions, 5. Run or proxy), since no Step 2 heading exists.
Remove the marketing paragraph about Appmachine and the filler line "Use action names and parameters as needed.", replacing the entity list with a one-line note that available actions are discovered via `membrane action list`.
Add validation guidance for data-mutating actions (e.g., read back the record after write, or check the `output` field and error states before batch changes) to lift workflow clarity above the destructive-operation cap of 3.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is mostly efficient commands and tables, but includes unnecessary padding: the marketing overview "Appmachine is a platform that allows users to create native and web apps without coding. It's primarily used by small businesses...", the filler line "Use action names and parameters as needed.", and a duplicate action-list snippet in 'Popular actions'. This matches 'Mostly efficient but includes some unnecessary explanation or could be tightened', above 2 because the core guidance is unpadded. | 3 / 5 |
Actionability | Concrete executable commands throughout: `npm install -g @membranehq/cli@latest`, `membrane login complete <code>`, `membrane action run <actionId> --connectionId=... --input '{"key": "value"}'`, plus a concrete flag table for proxy requests. Minor gaps (redundant action-discovery snippets, placeholder values like CONNECTION_ID never resolved) keep it below 5. | 4 / 5 |
Workflow Clarity | The connection workflow has explicit validation (poll with `--wait` until state changes, explicit handling of READY/CLIENT_ACTION_REQUIRED/CONFIGURATION_ERROR with follow-up actions), but coherence breaks elsewhere: "skip to **Step 2**" references a step that does not exist (only "1b" is numbered), the entity-list overview ends with the vague directive "Use action names and parameters as needed", and running data-mutating actions has no verification guidance — the destructive/batch cap applies. Steps exist but checkpoints are missing or the sequence is broken, matching anchor 3. | 3 / 5 |
Progressive Disclosure | No bundle files exist, so this is a single-file skill with clear section headers (Install, Authentication, Connecting, Searching, Running actions, Proxy, Best practices) matching 'Good structure; most content is appropriately placed'. Not 5 because at ~145 lines the proxy flags table and agent-type details could be split into a reference file, and the under-50-lines exception does not apply. | 4 / 5 |
Total | 14 / 20 Passed |