Content
92%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 instruction skill: every section is an actionable rule, sequence, or limit, capped by an explicit Quality Gate validation checklist with a delete-and-rewrite remediation path, and no bundle files are needed at this size. The only gap is the absence of a worked example email to anchor the intended register and length of the final output.
Suggestions
Add one short annotated example cold email (or a filled-in template for the five-part Cold Email Structure) so the target register, length, and ask phrasing are concrete rather than inferred.
Include a mini example of the under-100-word forwardable blurb for warm intro requests, since the word cap is stated but the format is not modeled.
Show one 'before → after' rewrite of a banned phrase (e.g. 'I'd love to connect' → a specific tie-in opener) to make the Hard Bans remediation demonstrable, not just declarative.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is lean checklists and short structural rules — e.g. the Cold Email Structure section is five one-line steps — with zero padding and no explanation of concepts Claude already knows (no 'what a cold email is', no fundraising 101). Every token earns its place, matching the level-5 anchor. | 5 / 5 |
Actionability | Guidance is concrete throughout: a numbered five-part cold email structure, an exact follow-up cadence ('day 0', 'day 4 or 5', 'day 10 to 12'), a hard 100-word cap on forwardable blurbs, and a literal banned-phrase list. It falls short of the level-5 anchor ('specific examples cover the common cases') only because no worked example or template of an actual email is provided, leaving the exact register of a finished draft implicit. | 4 / 5 |
Workflow Clarity | Each message type has a clearly sequenced structure, and the Quality Gate is an explicit pre-delivery validation checklist ('the ask is explicit', 'the proof point is concrete', 'filler praise and softener language are gone') with a defined remediation path — Hard Bans instructs to 'Delete and rewrite'. No destructive or batch operations exist that would cap the score. This matches the level-5 anchor's sequence-plus-checklist pattern. | 5 / 5 |
Progressive Disclosure | No bundle files exist, and none are needed: the body is a compact set of well-organized sections (When to Activate, Core Rules, Hard Bans, per-message-type structures), none of which is bulky enough to warrant extraction. Per the rubric's guideline for short self-contained skills, well-organized sections with no external-reference need score 5. | 5 / 5 |
Total | 19 / 20 Passed |