Content
71%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 solid, command-first integration skill: concrete CLI examples, explicit connection-state handling, and sensible best practices. The main defects are a dangling 'Step 2' reference, a 'Popular actions' section that lists none, filler lines, and reference material (entity catalog, proxy options) that is inlined but too thin to be useful.
Suggestions
Fix the workflow numbering: define the steps after 'Connecting to Agiloft' (or renumber '1b') so the repeated 'skip to **Step 2**' references resolve.
Make 'Popular actions' deliver on its title by naming 3–5 concrete Agiloft action ids with example invocations, or remove the section and fold the discovery command into 'Running actions'.
Either expand the 'Agiloft Overview' entity list with what each entity means for action discovery, or move it (with the proxy flag table) into a reference file and link to it from SKILL.md; also delete the filler line 'Use action names and parameters as needed.'
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is command-driven with little explanation of concepts Claude already knows, but contains trimmable filler such as 'Use action names and parameters as needed.', the cryptic entity list under 'Agiloft Overview', and promotional prose like 'This is the fastest way to get a connection.' This matches anchor 4 ('Efficient; minor instances of over-explanation that could be trimmed') rather than 5, where every token earns its place. | 4 / 5 |
Actionability | Concrete, executable CLI commands appear throughout — install (`npm install -g @membranehq/cli@latest`), login, `membrane connection ensure`, `action list`/`action run` with `--input` JSON examples, and a proxy flags table. Minor gaps keep it below anchor 5: `CONNECTION_ID` placeholders are never concretized and 'Popular actions' names no actual action ids. | 4 / 5 |
Workflow Clarity | The install → login → connection ensure → poll → discover → run sequence is clear, and the connection lifecycle includes a genuine feedback loop with explicit state handling (READY, CLIENT_ACTION_REQUIRED, CONFIGURATION_ERROR) and re-polling. It falls short of anchor 5 because 'skip to **Step 2**' references a step that is never defined (only '1b' exists) and 'Popular actions' is an empty pointer rather than a step. | 4 / 5 |
Progressive Disclosure | There are no bundle files (references/, scripts/, assets/ are absent) and the ~140-line body has reasonable section headers, but the entity catalog ('Knowledge Base / Article / Record...') and the proxy flag reference sit inline and underspecified — material that would serve better as a separate reference file or a fleshed-out section. This fits anchor 3 ('Some structure but could be better organized') rather than 4. | 3 / 5 |
Total | 15 / 20 Passed |