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 is highly actionable with complete CLI commands and a sensible connect-then-run workflow including state-based error handling, but it has a broken internal step reference ("skip to Step 2"), no verification guidance for the destructive delete actions it advertises, and some filler text. The single-file structure is well organized though it inlines a sizable action reference table.
Suggestions
Fix the dangling "skip to **Step 2**" reference by actually labeling the subsequent steps (Step 2: search actions, Step 3: run actions) or removing the step numbering entirely.
Add a validation checkpoint before destructive actions (delete-project, delete-card), e.g. confirm the entity via get-project/get-card and re-confirm with the user before running the delete.
Trim filler ("Use action names and parameters as needed") and either deepen the shallow Overview section with the actual entity relationships or remove it, moving the popular-actions table to a references file.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is mostly command examples and tables, but includes unnecessary filler such as "Use action names and parameters as needed", a shallow Overview section, and wordy asides ("This is the fastest way to get a connection..."), matching 'mostly efficient but could be tightened'. Not 4 because several padding spots exist, not 2 because there is no extensive explanation of concepts Claude already knows. | 3 / 5 |
Actionability | Fully executable copy-paste commands cover every phase: `npm install -g @membranehq/cli@latest`, `membrane login`, `membrane connection ensure`, `membrane action list --intent`, `membrane action run --input`, and the proxy flag table. Not 4 because the examples are complete and cover the common cases, including parameter passing and JSON output. | 5 / 5 |
Workflow Clarity | The connect-search-run sequence exists with state-based polling and error-state handling, but "skip to **Step 2**" references a step that is never labeled in the document (only "1b" exists), and destructive actions (delete-project, delete-card) have no validation/verification guidance, capping this at 3. Not 2 because the connection workflow does include explicit checkpoint handling (READY, CLIENT_ACTION_REQUIRED, error states). | 3 / 5 |
Progressive Disclosure | A single SKILL.md (no bundle files) with well-organized sections covering setup through advanced usage; the ~20-row actions table and proxy flag reference are borderline content that could live in a separate reference file. Not 5 because the file is well over 50 lines and carries reference-style material inline; not 3 because the structure is clear and easy to navigate with no buried references. | 4 / 5 |
Total | 15 / 20 Passed |