Content
40%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 strong on concrete, mostly executable code but weak on structure: it monolithically inlines standard testing patterns, offers only boilerplate workflow guidance with no validation checkpoints, and references bundle files that do not exist. Moving the inlined examples into the advertised (currently missing) reference files and replacing the generic Instructions with a sequenced workflow would fix most weaknesses.
Suggestions
Create the advertised bundle files (references/hardhat-setup.md, references/test-patterns.md, assets/test-suite.js, scripts/test-contract.sh, etc.) and move the long inlined code blocks into them, leaving SKILL.md as a lean overview with clearly signaled one-level-deep pointers.
Fix the broken and inconsistent reference paths — the Instructions point to 'resources/implementation-playbook.md' while the Resources section lists 'references/...' — so every referenced path resolves.
Replace the generic Instructions ('Clarify goals... Apply relevant best practices and validate outcomes') with a concrete sequenced workflow (setup config → write unit tests → fuzz edge cases → fork mainnet → run coverage) with explicit validation checkpoints such as 'only proceed when npx hardhat coverage passes and gas assertions hold'.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The ~430-line body inlines extensive standard patterns Claude already knows (basic loadFixture/Chai describe-it structure, vm.prank cheatcodes) plus boilerplate filler ('Clarify goals, constraints, and required inputs. Apply relevant best practices and validate outcomes.'). Several padded sections match the anchor 'noticeably verbose', and it does not reach anchor 3 because the padding is pervasive rather than occasional. | 2 / 5 |
Actionability | The body provides mostly executable guidance: a complete hardhat.config.js, a Foundry test contract with fuzz and fork tests, 'npx hardhat coverage', 'npx hardhat verify --network mainnet ...', and CI YAML. Minor gaps (undefined addr1/contractAddress variables in the impersonation and verification snippets, network never imported) keep it below the copy-paste-ready anchor 5. | 4 / 5 |
Workflow Clarity | There is no real sequence: the Instructions section is generic boilerplate ('Clarify goals... Apply relevant best practices and validate outcomes') with no concrete steps or validation checkpoints, and the remaining sections are topical reference material rather than an ordered workflow. It sits below anchor 3, whose example at least lists concrete sequenced steps. | 2 / 5 |
Progressive Disclosure | The body points to 'resources/implementation-playbook.md' and lists nine files under references/, assets/, and scripts/ — none of which exist in the bundle — while ~400 lines of detail that the Resources section claims live in those files are inlined in SKILL.md. Broken, inconsistently prefixed references ('resources/' vs 'references/') plus inlined detail match the anchor 'content that clearly belongs in separate files is inlined; or references are buried'. | 2 / 5 |
Total | 10 / 20 Passed |