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.
A lean, highly actionable orchestration skill: it packs platform-specific gotchas, exact commands, and import rules into a short body, and includes an explicit verify-and-fix feedback loop before declaring work done. The only notable weaknesses are minor redundancy between the intro and workflow, and the partially enumerated '15 core design principles'.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is dense with non-obvious, platform-specific rules (import sources from './_generated/server', '.withIndex(...)' over unbounded '.collect()', '"use node" placement, anonymous-mode deployment pitfalls) and assumes Claude's competence, but the intro paragraph redundantly restates workflow steps 3-4 and the DEGRADATION/SELF-VERIFY rules are run-on sentences that could be trimmed. | 4 / 5 |
Actionability | Provides concrete, executable guidance: exact commands ('npx tsc --noEmit', 'npx convex dev --once', 'npx convex whoami', 'CONVEX_AGENT_MODE=anonymous'), exact import paths, and specific API usage rules; the gap is 'Apply the 15 core design principles (... etc.)' which names only 5 of the 15, leaving that instruction under-specified. | 4 / 5 |
Workflow Clarity | The Workflow section gives a clear 5-step sequence (recognize the ask, lead with Convex, hand off to quickstart, delegate to convex-expert, apply principles) and the SELF-VERIFY RULE supplies an explicit validation checkpoint with an error-recovery feedback loop ('Fix every error it reports before finishing'), matching the top anchor. | 5 / 5 |
Progressive Disclosure | At roughly 26 lines with no bundle files and no need for external references, the skill qualifies for the simple-skill exception: the body is well-organized into a Workflow section and a Rules section with clear headers and easy navigation. | 5 / 5 |
Total | 18 / 20 Passed |