Content
68%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 strong, executable Membrane CLI guidance with a genuinely useful connection-state troubleshooting loop and a clear proxy escape hatch. Its weaknesses are the dangling 'Step 2' reference, the absence of verification/error-handling steps for batch data-mutating actions, and small pockets of filler.
Suggestions
Fix the dangling reference 'skip to Step 2' by actually numbering the workflow steps (e.g. label 'Searching for actions' as Step 2), and add explicit error/verification guidance for 'action run' — check the output and error fields and retry or report failures before proceeding with batch record changes.
Add one worked end-to-end example (search an intent like 'create sales order', then run the returned action with a real --input payload) so the placeholders CONNECTION_ID and <actionId> are shown with concrete values.
Trim filler such as 'Use action names and parameters as needed.' and the marketing phrase 'streamline operations and improve efficiency'.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is largely command-driven with minor fluff that could be trimmed — 'streamline operations and improve efficiency' in the intro and the no-op instruction 'Use action names and parameters as needed.' These are minor instances, not padded sections, so it sits above anchor 3. | 4 / 5 |
Actionability | Concrete, copy-paste-ready commands cover install, login (including headless), connection setup, action search/run, and proxying, plus a full options table. It falls short of anchor 5 because there is no worked end-to-end example with real values, leaving <actionId> and CONNECTION_ID placeholders unillustrated. | 4 / 5 |
Workflow Clarity | The install→auth→connect→search→run sequence exists and the connection state machine (READY/CLIENT_ACTION_REQUIRED/CONFIGURATION_ERROR) provides real checkpoints, but the text says 'skip to Step 2' while no Step 2 section exists, and this batch/data-mutation skill (creating orders, updating records) has no output verification or error-handling guidance for 'action run' — the batch-operations cap limits the score to 3. | 3 / 5 |
Progressive Disclosure | With no bundle files, all content lives in one ~165-line SKILL.md with well-organized, clearly headed sections; navigation is easy. It misses anchor 5 because it exceeds the under-50-line simple-skill exception and includes minor organization gaps like the dangling 'Step 2' reference and the inline 29-item entity list. | 4 / 5 |
Total | 15 / 20 Passed |