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 body is a well-structured, highly actionable reference for Canva automation via Rube MCP, with exact tool names, parameters, and async-polling checkpoints throughout. Its main weaknesses are duplicated pitfall guidance across sections and missing error-recovery loops for failed jobs.
Suggestions
Deduplicate the "Known Pitfalls" section and per-workflow "Pitfalls" lists — state each pitfall once (async polling and pagination currently appear multiple times) and reference it from workflows.
Add explicit failure handling for async jobs: what to do when status is 'failed' (retry, re-initiate the job, surface the error), turning the poll loop into a validate-recover feedback loop.
State the "call RUBE_SEARCH_TOOLS first" rule once (e.g., in Prerequisites) instead of repeating it in the description, Prerequisites, and body.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is mostly dense, useful reference material (tool slugs, parameters, enums), but there is real duplication: "Known Pitfalls" restates per-workflow pitfalls (async polling appears in workflows 3, 4, and 6 and again in Known Pitfalls; pagination appears twice), the Quick Reference table re-lists every tool and its params, and "Always call RUBE_SEARCH_TOOLS first" is stated three times. This fits 'mostly efficient but could be tightened'; not 2 because nothing explains concepts Claude already knows and no section is pure padding. | 3 / 5 |
Actionability | Concrete, executable guidance throughout: exact tool slugs (CANVA_CREATE_ASSET_UPLOAD_JOB), parameter names with accepted values ("format: 'pdf', 'png', 'jpg', 'svg', 'mp4', 'gif', 'pptx'"), status enums ('in_progress', 'success', 'failed'), and a specific 2-3 second poll interval. Not 5 because there are no example tool-call payloads showing arguments in practice, leaving minor gaps. | 4 / 5 |
Workflow Clarity | Each workflow has a numbered tool sequence, and async operations include explicit checkpoints (poll until 'success'/'failed', "Confirm connection status shows ACTIVE before running any workflows"), matching 'clear sequence with most checkpoints present'. Not 5 because there is no error-recovery guidance for what to do when a job returns 'failed' (no retry/backoff loop despite mentioning exponential backoff for rate limits); the cap-3 rule for destructive/batch operations does not apply since these operations are neither. | 4 / 5 |
Progressive Disclosure | No bundle files exist and the skill is a single self-contained document; structure is good — clear section headers, per-workflow organization, and a navigable Quick Reference table, with no buried or nested references. This fits 'good structure; most content is appropriately placed; minor organization gaps'. Not 5 because at ~215 lines the per-workflow parameter and pitfall detail could arguably be split into reference files, and not 3 because navigation is easy and nothing is misfiled. | 4 / 5 |
Total | 15 / 20 Passed |