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.
An excellent traps-oriented reference: terse, symptom-first sections with mostly copy-paste-ready commands and well-signaled pointers to the wiki for bulk detail. The two weaker spots are fixes described prose-only (lazy thunks, routes TOML) and missing post-change verification in the secret workflows.
Suggestions
Add a complete copy-paste example for the two prose-only fixes: a lazy-thunk/request-time snippet for the module-evaluation crash and a full `[[env.production.routes]]` TOML block for the named-environment routes trap.
Add a verification step after secret changes (e.g., confirm with `wrangler secret list` or a test request) and after the forced redeploy, since section 3 notes warm isolates keep old behaviour.
Section 5 could name where the routes block goes (top level vs `env.production` in which file) to remove the last ambiguity for a first-time fixer.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is lean and assumes competence: no explanation of what Cloudflare Workers or wrangler are, and every sentence delivers either a trap, a command, or a pointer ("This skill holds the traps that cost time"). It matches the score-5 anchor — every token earns its place — rather than score 4, which would require trimmable over-explanation. | 5 / 5 |
Actionability | Most guidance is copy-paste ready: `bunx wrangler tail <worker-name> --env production`, the grep/sed/printf secret pipe, `gh workflow run production-release.yml -f surfaces=osn_api -f force=true`, and `bun run test`. It falls short of score 5 because two fixes are described without complete code — the module-evaluation fix ("Move both into request-time code or lazy thunks") and the routes fix name `[[env.production.routes]]` without a full TOML block. | 4 / 5 |
Workflow Clarity | Each section is a clearly sequenced conditional procedure keyed to a symptom ("When a request ... misbehaves, run `wrangler tail` ... before forming any theory"), and section 4 includes an explicit verification checkpoint ("Verify with a real `wrangler deploy`; `--dry-run` does not catch it"). It sits below score 5 because the secret-setting and redeploy flows lack a post-change verification step (e.g., confirming the secret landed or the behaviour changed), which matters since `wrangler secret put/delete` is a risky operation. | 4 / 5 |
Progressive Disclosure | The body is under 50 lines with well-organized numbered sections, and bulk material (tier map, secrets checklist, deploy steps) is pointed out cleanly to `wiki/shared/production-deploy.md` and `wiki/shared/dev-environment.md` — one level deep, clearly signaled, no nesting. No bundle files exist to misorganize, so this matches the simple-skill case for a top score. | 5 / 5 |
Total | 18 / 20 Passed |