Content
76%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 strongly actionable, well-structured single-file skill: exact CLI commands, a comprehensive action catalog, and a clear connection workflow with state-based checkpoints. The main gap is the absence of validation before destructive actions (delete-incident/heartbeat/monitor), which caps workflow clarity, plus minor padding in the overview and best-practices sections.
Suggestions
Add explicit validation before destructive actions, e.g. "Before running delete-incident, fetch it with get-incident and confirm the ID and title with the user; deletion is permanent."
Trim the opening paragraph about what Better Stack is and who uses it — Claude already knows this — and drop marketing framing like "so you can focus on the integration logic rather than auth plumbing".
Clarify the relationship between the action table's "Key" column and the <actionId> used in `membrane action run`, e.g. show one fully concrete example such as `membrane action run resolve-incident --connectionId=... --input '{"incidentId": "..."}'`.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is largely command-focused with copy-paste CLI invocations, but has minor trimmable padding: the opening paragraph explaining what Better Stack is and who uses it ("It's used by DevOps engineers and SREs to monitor their applications"), and framing like "so you can focus on the integration logic rather than auth plumbing" and "This will burn less tokens". This matches 'efficient; minor instances of over-explanation that could be trimmed' rather than the fully lean anchor. | 4 / 5 |
Actionability | Every section provides exact, copy-paste-ready commands with flags and JSON examples — "npm install -g @membranehq/cli@latest", "membrane connection ensure \"https://betterstack.com/\" --json", "membrane action run <actionId> --connectionId=CONNECTION_ID --input '{\"key\": \"value\"}' --json" — plus a 16-row action table with keys and a proxy options table, fully covering the common cases. | 5 / 5 |
Workflow Clarity | The install → login → connection ensure → poll state → search → run sequence is well laid out with state-based checkpoints (READY / BUILDING / CLIENT_ACTION_REQUIRED with follow-up polling, and error handling for CONFIGURATION_ERROR / SETUP_FAILED). However, the skill drives destructive operations (delete-incident, delete-heartbeat, delete-monitor) with no validation or verification steps before executing them, and the rubric explicitly caps workflow clarity at 3 for destructive operations without validation. | 3 / 5 |
Progressive Disclosure | As a single-file skill with no bundle files, the content is organized into well-labeled sections (Install, Authentication, Connecting, Searching, Popular actions, Running, Proxy, Best practices) with no buried or nested references, and the action table is appropriately core inline content — matching 'good structure; most content is appropriately placed' with only minor gaps (e.g., the proxy flag table and best-practices sections could be tightened into fewer places). | 4 / 5 |
Total | 16 / 20 Passed |