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.
A highly actionable, well-sequenced operational manual with strong validation checkpoints and feedback loops. Its weaknesses are repetition that inflates token cost and a monolithic structure with no progressive disclosure into bundle files despite its length.
Suggestions
Deduplicate repeated command blocks (env status --json, env add/use, managed-instance create/start, 'stop after two repeated failures') by defining each once and cross-referencing, to tighten conciseness from 3 toward 4-5.
Extract large self-contained sections (e.g. 'Optional Native n8n MCP Assist' routing policy, AI/LangChain node rules, full Knowledge Commands list) into reference files under ./references/ and link one level deep, improving progressive_disclosure from 3.
Consolidate the repeated secret-handling rules ('do not print API keys', 'prefer --api-key-stdin', 'do not pass secrets inline') into a single 'Secret Handling' section referenced where needed.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is operationally dense with no concept padding, but it is ~520 lines with notable repetition (env status --json, env add blocks, managed-instance create/start commands, 'do not print API keys', and 'stop after two repeated failures' each appear multiple times), placing it at 'mostly efficient but could be tightened' rather than the minor-trim level of 4. | 3 / 5 |
Actionability | Copy-paste-ready `npx --yes n8nac` commands, concrete TypeScript decorator examples, and specific flags (--api-key-stdin, --verify, --prod, --query) cover the common cases; placeholders like 'exact type from node-info' are intentionally parameterized by the schema-first rule, not pseudocode. | 5 / 5 |
Workflow Clarity | Explicit numbered Bootstrap Order (8 steps) and Operating Loop (11 steps) with validation checkpoints (pull-before-edit, push --verify, conflict resolution modes, 'stop after two repeated failures') provide clear sequencing and feedback loops for destructive/batch operations. | 5 / 5 |
Progressive Disclosure | Section headers give real structure, but no references/scripts/assets bundle files exist, the large 'Optional Native n8n MCP Assist' routing section and AI/LangChain rules are inlined in a ~520-line monolithic SKILL.md, and nothing points one level deep to separate files, matching 'some structure but content that should be separate is inline'. | 3 / 5 |
Total | 16 / 20 Passed |