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 strong, executable guide: real CLI commands with JSON flags, a connection lifecycle with state-by-state handling, and a proxy escape hatch with a complete options table. Its weaknesses are moderate redundancy and filler, an empty 'Popular actions' section, and broken step numbering ('Step 2' and '1b' references without matching headings).
Suggestions
Merge 'Searching for actions' and 'Popular actions' into one section, settle on a single invocation form (`membrane` vs `npx @membranehq/cli@latest`), and populate 'Popular actions' with a few real example actions (e.g. listing or creating an invoice) with sample inputs.
Fix the step numbering: replace 'skip to Step 2' and the orphan '1b' heading with a consistent labeled sequence (Step 1: connect, Step 2: discover actions, Step 3: run them), and add a brief note on interpreting errors from `action run` output.
Trim filler lines ('Use action names and parameters as needed', the 'Agent Types ... etc.' aside) and move the connection-state reference (clientAction fields, state meanings) into a separate reference file to keep the main flow lean.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | Most of the body is efficient, concrete command documentation, but there is redundancy (the action-list command appears in both 'Searching for actions' and 'Popular actions' with inconsistent forms) and filler such as 'Use action names and parameters as needed' and the vague 'Agent Types : claude, openclaw, codex, warp, windsurf, etc.' line. This matches 'Mostly efficient but includes some unnecessary explanation or could be tightened' rather than level 4, which would have only minor trimmable instances. | 3 / 5 |
Actionability | The body provides copy-paste-ready executable commands for every phase (install, login, connection ensure, polling, action list/run with JSON input, proxy requests with a flags table), matching 'Mostly executable guidance; concrete code or commands with minor gaps'. It falls short of level 5 because the 'Popular actions' section lists no actual actions and no complete end-to-end example (e.g. a real action run with its input and output) is shown. | 4 / 5 |
Workflow Clarity | The setup sequence is clear and includes checkpoints: poll with --wait, interpret state (READY, CLIENT_ACTION_REQUIRED, CONFIGURATION_ERROR/SETUP_FAILED), and act on clientAction fields before re-polling. It misses level 5 because the text says 'skip to Step 2' and uses a '1b' heading without any labeled Step 1/Step 2, leaving the step numbering incoherent, and there is no error-recovery guidance for action runs themselves. | 4 / 5 |
Progressive Disclosure | There are no bundle files, and the body is well-organized into clearly headed sections (install, authentication, connecting, searching, running, proxy, best practices) with most content appropriately placed inline for a ~145-line skill. This matches 'Good structure; most content is appropriately placed; minor organization gaps' — a reference file holding an action catalog or connection-state details would improve it, keeping it below level 5. | 4 / 5 |
Total | 15 / 20 Passed |