Content
77%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.
Highly actionable, tool-accurate reference material with exemplary validation loops and safety gates. It is held back by redundant restatement across sections and a guide-file split that either is missing from the bundle or is undercut by deep inline detail in SKILL.md itself.
Suggestions
Merge the 'Quick Reference' speed column and the 'Performance Characteristics' table into one table, and cut the 'Summary' section's restatement of Common Mistakes and patterns to pure pointers.
State the nodeType short-vs-full prefix rule once (in 'Critical: nodeType Formats') and reference it from Common Mistakes, the Best Practices lists, and the Summary instead of repeating it in each.
Ship the four referenced guide files with the skill, and move the deep per-tool detail (n8n_test_workflow methods, n8n_workflow_versions sources, n8n_manage_agents actions/error codes) into them, leaving one-paragraph summaries plus links in SKILL.md.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is dense with server-specific facts Claude cannot know (good), but duplication is pervasive: the Quick Reference 'Speed' column restates the Performance Characteristics table, the Summary restates Common Mistakes and patterns, nodeType-format guidance appears in four places (Tool Selection, Critical: nodeType Formats, Common Mistakes #1, Summary #2), and timing stats like '18s average' and '56s avg between edits' repeat more than four times. | 3 / 5 |
Actionability | Fully executable throughout: concrete calls like `search_nodes({query: "slack"})` and `n8n_manage_credentials({action: "list"})`, wrong/correct JSON for credentials and nodeType prefixes, exact parameter shapes ('column: {name, type}', '{__rl: true, mode: "id", value: …}'), and named error codes per tool. | 5 / 5 |
Workflow Clarity | Multi-step processes are explicitly sequenced with validation checkpoints and feedback loops ('Fix errors, validate again', 'Validate after every significant change', validate → fix → validate cycles), and destructive or batch operations get explicit confirm/dryRun gates ('confirm with the user before dropping a populated column', 'use dryRun: true before bulk changes', 'ask the user first' for exposeToMcp). | 5 / 5 |
Progressive Disclosure | References to SEARCH_GUIDE.md, VALIDATION_GUIDE.md, WORKFLOW_GUIDE.md, and OPERATIONS_GUIDE.md are clearly signaled one level deep, but none of those files exist in the bundle, and long inline sections (Running Workflows, Version History, Agents) carry detail that the guide split was meant to offload — structure exists, but content placement and broken navigation hold it at the middle anchor. | 3 / 5 |
Total | 16 / 20 Passed |