Content
90%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.
This is a tight, expert-level instruction skill: every section carries n8n-3.0-specific knowledge, the guidance is executable down to exact parameter names and expressions, and the process sequence is clear. The only gaps are minor: the save-time validation loop is not an explicit final Process step, and the node-replacement catalog could be split into a reference file.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is dense with n8n-version-specific facts Claude cannot know (mode mappings, parameter renames, expression rewrites) and contains zero padding or explanation of general concepts; every line earns its place, matching the lean-and-efficient anchor. | 5 / 5 |
Actionability | Guidance is copy-paste concrete throughout: exact mode names ("Run Once for All Items"), exact rewrites (`item` to `item.json`, `getWorkflowStaticData(type)` to `$getWorkflowStaticData(type)`), exact expression syntax (`$fromAI('<name>', '<description>', '<type>')`, `$('<node name>').itemMatching(index)`), and per-node replacement recipes covering the common cases. | 5 / 5 |
Workflow Clarity | The three-step Process is clearly sequenced with a decision checkpoint ("ask one focused question with ask_user before you edit") and the report's save-time re-check is stated as the built-in verification ("re-checks a workflow each time it is saved, so a correct fix closes the finding by itself"). It falls short of the 5 anchor because an explicit 'save and confirm the finding closes' step is never integrated into the Process sequence itself — the validation loop is described as context rather than as a step. | 4 / 5 |
Progressive Disclosure | The skill is a single well-sectioned file with clear headers (Scope, Process, Node replacements, AI Agent modes, Behavior changes) and no nested references, which suits its moderate size. It stays at 4 rather than 5 because the ~60-line node-replacement catalog is reference material that could arguably live in a one-level-deep references/ file, leaving SKILL.md as a leaner overview. | 4 / 5 |
Total | 18 / 20 Passed |