Content
85%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 highly actionable and workflow-sequenced with strong validation discipline, assuming Claude's competence throughout. Its main weaknesses are repeated command blocks across sections and a single-file monolith that would benefit from splitting detailed reference material into bundled files.
Suggestions
Split the Knowledge Commands catalog, AI/LangChain Node Rules, and Common Mistakes into separate reference files under ./references/ and link to them one level deep, leaving SKILL.md as a concise overview — this would raise progressive_disclosure and trim the monolith.
De-duplicate repeated command blocks: `env status --json`, `env add ... --workflows-path`, and the managed-instance `create`/`start` sequences each appear in multiple sections — reference one canonical block instead of restating.
Consolidate the Bootstrap Order, Environments, and Managed Local Runtime sections so the readiness decision tree is stated once, reducing token cost without losing the validation checkpoints.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is lean and assumes Claude's competence (no explanations of what n8n or TypeScript decorators are), but several commands are repeated across sections — `npx --yes n8nac env status --json`, `env add`, and the managed-instance `create`/`start` blocks appear two or three times — so not every token earns its place. | 4 / 5 |
Actionability | It provides copy-paste-ready bash commands and a complete executable TypeScript workflow skeleton, plus exact parameter shapes such as `{ __rl: true, value: '...', mode: 'list' | 'id' | 'url' }`, covering the common authoring and sync cases. | 5 / 5 |
Workflow Clarity | The Bootstrap Order and Operating Loop are numbered sequences with explicit validation checkpoints (`accessStatus: "ready"`, `push --verify`, `test-plan`), and destructive/batch operations have feedback loops (stop-and-inspect on conflict, stop after two repeated failures), so the destructive cap does not apply. | 5 / 5 |
Progressive Disclosure | The file is well-sectioned with clear headers, but it is a ~530-line monolith with no bundle files and no one-level-deep references; content such as the long Knowledge Commands catalog, AI/LangChain rules, and Common Mistakes could be split into separate reference files for easier navigation. | 3 / 5 |
Total | 17 / 20 Passed |