Content
67%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 content is a well-structured, action-oriented guide with concrete CLI commands, a useful action catalog, and a strongly validated connection workflow. Its main weaknesses are an unnecessary introductory explainer of what Salesforce is and the absence of verification guidance around the destructive/batch actions the skill surfaces.
Suggestions
Remove or drastically condense the opening "Salesforce is a leading cloud-based CRM platform..." paragraph; assume Claude already knows what Salesforce is.
Add an explicit validation/verification checkpoint to the "Running actions" workflow for destructive and batch operations (e.g., confirm before delete-multiple-records, verify the returned record count or IDs after batch create/update).
Consider moving the 20-row "Popular actions" catalog and the proxy flag table into a separate reference file referenced one level deep from SKILL.md to reduce inline bulk.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is mostly efficient concrete commands and tables, but the opening paragraph ("Salesforce is a leading cloud-based CRM platform that helps businesses...") explains a concept Claude already knows, and phrases like "so you can focus on the integration logic rather than auth plumbing" add padding the rubric explicitly penalizes. | 3 / 5 |
Actionability | It provides many concrete, executable commands (npm install, membrane login, connection ensure, action list/run, request) plus a populated actions table, but execution depends on a prior discovery step and uses placeholders like <actionId> and CONNECTION_ID, leaving minor gaps versus copy-paste-ready coverage. | 4 / 5 |
Workflow Clarity | The connection lifecycle is a clear sequence with explicit validation and feedback loops (poll --wait, READY/BUILDING/CLIENT_ACTION_REQUIRED/CONFIGURATION_ERROR handling), but the generic "Running actions" step offers no verification checkpoint for the destructive/batch actions (delete-multiple-records, create up to 200) it exposes, which is a notable gap. | 4 / 5 |
Progressive Disclosure | The single SKILL.md is well-organized with clear section headers and easy navigation, and there are no nested references; however at ~185 lines with a sizable inline actions catalog and proxy-options table, some of this bulk could live in one-level-deep reference files rather than entirely inline. | 4 / 5 |
Total | 15 / 20 Passed |