Conciseness

The bulk of the body is command-dense and efficient, but it carries several unnecessary elements: the 'Anymail Finder Overview' stub tree ('Search / Search Result / Account') with no content, a marketing-style background sentence ('It's used by sales teams, marketers, and recruiters'), six 'Popular actions' rows with an empty Description column, and a rambling 'Agent Types' aside. This fits anchor 3 ('mostly efficient but includes some unnecessary explanation') better than 4, where over-explanation would be isolated and minor.

Actionability

Concrete, copy-paste-ready commands cover every phase: 'npm install -g @membranehq/cli@latest', 'membrane connection ensure "https://anymailfinder.com/" --json', 'membrane action run <actionId> --connectionId=CONNECTION_ID --input '{"key": "value"}' --json', plus a flag table for proxy requests. Minor gaps keep it at anchor 4 rather than 5: the 'Popular actions' table's empty Description column leaves the agent without parameter details (mitigated only by the pointer to 'inputSchema' from action list), and 'membrane login --tenant --clientName=<agentType>' is never explained.

Workflow Clarity

The setup sequence (install → login → connection ensure → poll states → action list → action run) is clearly laid out with real checkpoints: '--wait' long-polling until state leaves BUILDING, explicit handling of CLIENT_ACTION_ERROR states with clientAction fields, and re-polling after user action. It is not 5 because 'If the returned connection has state: "READY", skip to Step 2' references a 'Step 2' that is never defined anywhere in the document, breaking the sequence's navigation.

Progressive Disclosure

The body (~145 lines, no bundle files present) is organized into clearly headed sections (Install, Authentication, Connecting, Searching, Running, Proxy, Best practices) with a single one-level-deep external pointer (official docs URL) and no nested references. It is not 5 because the ~145-line single file inlines reference-style material (the full proxy flag table and the actions table) that could live in separate files, and the empty overview tree is not a useful navigation aid.

Specificity

The description names the domain ('Anymail Finder integration') but the actions are pure filler — 'Manage data, records, and automate workflows' describes virtually any integration and never mentions the actual capabilities (finding, verifying, or enriching email addresses). It sits at anchor 2 ('names the domain but actions are minimal or generic'), not 3, because no concrete domain-specific action is stated.

Completeness

The 'when' is explicit ('Use when the user wants to interact with Anymail Finder data') but circular — it only triggers for users who already know the tool — and the 'what' is vague ('Manage data, records, and automate workflows'), so neither side is fully strong. This matches anchor 3's one-good-one-weak profile rather than anchor 4, which requires both 'what' and 'when' to be present with only the 'when' improvable.

Trigger Term Quality

'Anymail Finder' and 'data' are the only keywords; the natural phrases a user would actually say — 'find email', 'verify email', 'email address', 'contact', 'lead' — are entirely absent, so common variations and synonyms are missing (anchor 3, e.g. 'Works with PDF files'). It is above anchor 2 because the app name itself is a genuinely relevant, non-generic keyword.

Distinctiveness Conflict Risk

'Anymail Finder' is a distinct niche app name, so the trigger is unlikely to fire for unrelated skills, matching anchor 4 ('mostly distinct; minor overlap risk'). Not 5 because the generic middle sentence ('Manage data, records, and automate workflows') overlaps with any data-management or automation skill's trigger space.