Content
100%Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.
The body is concise, fully actionable with runnable commands and code, and presents a well-sequenced workflow with explicit verification feedback loops. Its single bundle file is cleanly signaled and one level deep, giving good progressive disclosure.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is lean: it states the convention crisply and uses minimal Before/After TS snippets without explaining concepts Claude already knows (no "what is null vs undefined" padding), matching the score-3 lean-and-efficient anchor over the mostly-efficient score-2. | 3 / 3 |
Actionability | It provides an executable scanner command, copy-paste Before/After TS examples, concrete call-site patterns (`field: null`, `field: maybeValue ?? null`), and a runnable verify command (`bunx turbo run make --filter='...'`), matching the score-3 fully-executable anchor rather than the pseudocode/gap score-2. | 3 / 3 |
Workflow Clarity | The 7-step sequence includes an explicit Verify step with a feedback loop ("Run the scanner again until only intentional public API exceptions remain") plus focused tests/builds, so the batch-operation verification cap does not apply; this matches the score-3 checkpoint-and-recovery anchor over score-2's missing checkpoints. | 3 / 3 |
Progressive Disclosure | Content is organized into Rule / Workflow / Review checklist with one well-signaled, one-level-deep bundle reference (`scripts/find-new-optional-params.ts`) that exists on disk, matching the score-3 clear-overview-with-signaled-references anchor rather than the score-2 poorly-signaled case. | 3 / 3 |
Total | 12 / 12 Passed |