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 largely actionable, well-sectioned guide to driving BuiltWith via the Membrane CLI, with copy-paste commands, a clear connection lifecycle with state-based validation, and sensible best practices. Main weaknesses are padded introductory/overview sections, dangling "Step 2"/"1b" references that break the workflow's numbering, and a minor ambiguity between action "Key" and "<actionId>".
Suggestions
Cut the opening "BuiltWith is a web profiling tool..." paragraph and the nested "BuiltWith Overview" bullet list (it duplicates the Popular actions table), keeping one sentence of purpose and the capability map.
Fix the workflow numbering: either label the actual steps (1. Install & login, 2. Connect, 3. Find actions, 4. Run) so "skip to Step 2" and "1b" resolve, or reword the cross-references to section names.
Clarify that the "Key" column in the Popular actions table is the actionId used by "membrane action run", and add one complete worked example (e.g., get-domain-trust-score with a real domain).
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The bulk is lean (bash commands, tables, state-machine bullets), but there is noticeable padding: the opening paragraph explaining what BuiltWith is and who uses it ("BuiltWith is a web profiling tool that identifies the technologies used to build websites. Developers, researchers, and sales teams use it..."), the nested "BuiltWith Overview" bullet list that largely restates the Popular actions table, and filler like "Use action names and parameters as needed." This fits "Mostly efficient but includes some unnecessary explanation or could be tightened"; it is not a 4 because the padding spans several sections rather than minor instances. | 3 / 5 |
Actionability | Guidance is concrete and copy-paste ready throughout: "npm install -g @membranehq/cli@latest", "membrane login --tenant --clientName=<agentType>", "membrane connection ensure \"https://builtwith.com/\" --json", "membrane action run <actionId> --connectionId=CONNECTION_ID --input '{\"key\": \"value\"}' --json", plus a full proxy flag table. It falls short of 5 on minor gaps: the Popular actions table lists "Key" values while the run command takes "<actionId>" without stating they are the same, and there is no single worked end-to-end example with real values. | 4 / 5 |
Workflow Clarity | The sequence (install → login → connection ensure → poll state → search actions → run/proxy) is clear, with a genuine validation/feedback loop in the connection state machine (poll until READY, handle CLIENT_ACTION_REQUIRED by showing the user uiUrl/agentInstructions, then re-poll, and check error on CONFIGURATION_ERROR/SETUP_FAILED). It is not 5 because the flow references steps that don't exist as labeled sections — "skip to **Step 2**" and the heading "#### 1b." imply numbered steps 1/2 that are never defined — leaving the end-to-end path implicit, and error recovery beyond "check the error field" is thin. | 4 / 5 |
Progressive Disclosure | There are no bundle files (references/, scripts/, assets/ do not exist) and no external file references in the body, so there is no nested-reference problem; the single-file structure is appropriate for a CLI-integration skill of this size (~160 lines) and sections are clearly headed. It is not 5 because the doc exceeds the simple-skill threshold and content like the 10-row Popular actions table and proxy flag reference could be split into a separate reference file, though keeping them inline is defensible. | 4 / 5 |
Total | 15 / 20 Passed |