Content
57%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.
A generally actionable, command-driven skill body with concrete CLI examples and good state-handling guidance for connection setup. Weaknesses center on structure and safety: a broken 'Step 2' cross-reference, no validation guidance for destructive batch row operations, an empty descriptions column in the actions table, and no use of bundle files to keep the overview lean.
Suggestions
Fix the workflow numbering: define explicit Step 1/Step 2 headings or replace 'skip to Step 2' with the actual section name, and add verification guidance for batch-delete-rows/batch-create-rows (e.g., list rows to confirm IDs, or check the output count) before destructive runs.
Move the Popular actions table and the proxy request flag reference into a references/ file, keeping SKILL.md as a lean overview with one-level-deep pointers.
Trim the Airtable-comparison intro, the object hierarchy, and the filler line 'When to use which actions: Use action names and parameters as needed'; fill in the Description column of the actions table so action selection is self-explanatory.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The bulk is lean CLI commands and tables, but padding is present: a two-sentence "open-source alternative to Airtable" intro, a low-value Database/Table/Row/View object hierarchy, and the filler line "When to use which actions: Use action names and parameters as needed." — fitting 'mostly efficient but includes some unnecessary explanation or could be tightened'. | 3 / 5 |
Actionability | Commands are copy-paste ready throughout ("npm install -g @membranehq/cli@latest", "membrane action run <actionId> --connectionId=... --input '{...}' --json") with a full proxy flag table — mostly executable with minor gaps: the Popular actions table's Description column is entirely empty and CONNECTION_ID is never shown being captured concretely. | 4 / 5 |
Workflow Clarity | A rough sequence exists (install → login → connection ensure → poll state → search actions → run/proxy) with state-based checkpoints for BUILDING/CLIENT_ACTION_REQUIRED, but "skip to Step 2" references a Step 2 that is never defined, and the batch-delete/batch-create row actions — destructive batch operations — have no validation or verification guidance, capping this dimension at 3 per the rubric. | 3 / 5 |
Progressive Disclosure | No bundle files or external references exist; the ~157-line body inlines everything, including a 14-row actions table and a proxy flag reference that could live in separate files, and the broken 'Step 2' numbering hurts navigation — matching 'some structure but could be better organized; content that should be separate is inline'. | 3 / 5 |
Total | 13 / 20 Passed |