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 integration guidance with a well-checked connection lifecycle, but it never reaches AMLBot specifics — no named actions, no example wallet or transaction check — and duplicates/filler pad the token cost. Organization is decent for a single file, though generic connector material should be split into a reference. Fixing the dangling "Step 2" references and adding one concrete end-to-end AMLBot example would lift it substantially.
Suggestions
Delete the duplicate "Popular actions" section (it repeats the `action list` command verbatim) and either list 3–5 real AMLBot actionIds with example inputs, or fold it into "Searching for actions".
Fix the broken step numbering: the text says "skip to Step 2" twice but no Step 2 heading exists — either label the sections (Step 1: connect, Step 2: discover/run actions) or drop the references.
Move the generic Membrane proxy flag table and clientAction state semantics into a shared reference file (e.g. references/membrane-cli.md) and keep SKILL.md as the AMLBot-specific overview, which would also improve progressive disclosure.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is mostly command-driven, but the "Popular actions" section repeats the `action list` command already shown verbatim, the resource tree (Transaction/Account/Balance/...) adds no instruction, and the malformed "Agent Types : claude, openclaw..." sentence is padded filler. Matches anchor 3 (mostly efficient but could be tightened); not 4 because the duplication and filler are noticeable rather than minor. | 3 / 5 |
Actionability | Commands are concrete and executable throughout — install (`npm install -g @membranehq/cli@latest`), auth (`membrane login complete <code>`), connection (`membrane connection ensure "https://amlbot.com/" --json`), action execution with `--input` JSON, and a full proxy flag table. Matches anchor 4 (mostly executable, minor gaps); not 5 because no real AMLBot example (an actual actionId or wallet-check walkthrough) covers the common use case. | 4 / 5 |
Workflow Clarity | The sequence install → login → connection ensure → poll state → run action → proxy fallback is clear, and connection-state handling is an explicit checkpoint loop (poll while BUILDING, act on CLIENT_ACTION_REQUIRED then re-poll, check `error` on CONFIGURATION_ERROR/SETUP_FAILED). Matches anchor 4; not 5 because the text twice says "skip to Step 2" yet no Step 2 heading exists, and not 3 because the checkpoints are explicit rather than implicit. | 4 / 5 |
Progressive Disclosure | No bundle files exist (references/, scripts/, assets/ are absent) and everything is inline in a ~140-line file: the generic Membrane proxy flag table and clientAction semantics are reference-grade material that belongs in a separate file, and nothing signals where deeper detail lives. Matches anchor 3 (content that should be separate is inline); not 2 because sections are well organized and navigation within the file is easy. | 3 / 5 |
Total | 14 / 20 Passed |