Content
86%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 tight, high-signal debugging skill: every section is a non-obvious production trap paired with an executable command, and the bulk operational detail is correctly deferred to external wiki files. The only soft spots are the directive-without-code fix in section 4 and the absence of an explicit post-change verification step in the secret/redeploy flow.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is lean and trap-focused ('This skill holds the traps that cost time') with no padding or explanation of concepts Claude already knows; each of the six sections states a non-obvious failure mode and its fix in 2-4 lines. | 5 / 5 |
Actionability | Four of six sections give copy-paste-ready commands (`bunx wrangler tail`, the grep/sed/printf secret pipeline, `bunx wrangler deploy`, `bun run test`) plus a concrete routes snippet, but section 4 ('Move both into request-time code or lazy thunks') directs without showing code, and commands carry placeholders (worker-name, KEY, $SF) requiring substitution. | 4 / 5 |
Workflow Clarity | Numbered sections establish an explicit diagnostic order ('Tail the failing Worker first... before forming any theory') and validation is present ('Verify with a real `wrangler deploy`; `--dry-run` does not catch it'; 'needs the whole monorepo suite before merging'), but the piece is a traps catalog rather than one coherent multi-step workflow, and the production secret-change flow has no explicit post-deploy verification step. | 4 / 5 |
Progressive Disclosure | At ~43 lines with no bundle files, the skill keeps only the traps inline and clearly signals one-level-deep pointers up front ('The tier map, secrets checklist and deploy steps are in `wiki/shared/production-deploy.md` and `wiki/shared/dev-environment.md`'), with well-organized numbered sections. | 5 / 5 |
Total | 18 / 20 Passed |