Content
63%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 gives genuinely executable Membrane CLI workflows with strong state-handling and error-recovery guidance — its strongest aspect is actionability. Its weaknesses are a large unexplained inline entity list that wastes context, broken step numbering ('skip to Step 2' with no Step 2), and no progressive disclosure structure to offload detail.
Suggestions
Move the ~60-item entity list to a references file (e.g. 'See [ENTITIES.md](references/entities.md) for the full catalog of Asavie resources') or replace it with a one-line summary of the most common resource types.
Fix the step numbering: the text says 'skip to Step 2' but no Step 2 exists and the only numbered heading is '1b' — renumber the sections (1. Connect, 2. Wait for ready, 3. Discover actions, 4. Run) consistently.
Add one fully worked end-to-end example (search an action by intent, read its inputSchema, run it with --input, read the output field) so the most common path is copy-paste ready.
Trim filler such as 'Use action names and parameters as needed' and the multi-sentence product introduction, keeping only what changes how the agent executes.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The core CLI guidance is efficient, but the ~60-item bare entity list ('- **Profile** ... - **Shared Link**') carries no actionable information, and filler like 'Use action names and parameters as needed' plus introductory product prose add tokens without teaching anything Claude couldn't infer. This matches 'Mostly efficient but includes some unnecessary explanation or could be tightened'; it is not 4 because the entity list is a substantial padding block. | 3 / 5 |
Actionability | Commands are concrete and executable end-to-end: install command, login with headless fallback, 'membrane connection ensure "https://www.asavie.com/" --json', action list/run with '--input '{"key": "value"}', and a complete proxy flag table. This matches 'Mostly executable guidance; concrete code or commands with minor gaps' — the gaps being the dangling 'skip to Step 2' with no Step 2 and unexplained actionId sourcing. It is not 5 because common cases lack a fully worked example (e.g. one complete search-then-run sequence with real values). | 4 / 5 |
Workflow Clarity | There is a clear sequence (install → authenticate → connect → poll for readiness → discover actions → run), and the state-handling section provides genuine checkpoints: poll with '--wait', explicit handling of 'READY', 'CLIENT_ACTION_REQUIRED' with its clientAction fields, and 'CONFIGURATION_ERROR'/'SETUP_FAILED' error guidance. This matches 'Clear sequence with most checkpoints present; minor validation gaps' — the broken 'Step 2' reference, inconsistent numbering ('1b' with no '1'/'2'), and the undated auth-error retry loop keep it below 5. | 4 / 5 |
Progressive Disclosure | The body has reasonable section headers and no nested references, but the ~60-item entity inventory is exactly the kind of inline content that belongs in a separate reference file (or should be dropped entirely), and there are no bundle files at all to offload detail. This matches 'Some structure but could be better organized; content that should be separate is inline'. It is not 4 because a large low-value list is fully inlined rather than split out or summarized. | 3 / 5 |
Total | 14 / 20 Passed |