Content
71%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 strong, expert-calibrated strategy skill: concrete parameters, a clear pattern-selection decision framework, and an output template that makes the deliverable unambiguous. Main weaknesses are the absence of any progressive disclosure (everything lives in one long file despite several separable topics) and time-sensitive 2025 regulatory claims that are not quarantined, plus a handful of descriptive rather than instructive passages.
Suggestions
Move the Apple/Google compliance matrix and the Quiz Funnel Mechanics section into a one-level-deep reference file (e.g. references/compliance.md, references/quiz-mechanics.md) and keep SKILL.md as the pattern-selection overview with clearly signaled links, keeping the top-level file under ~100 lines.
Quarantine the dated regulatory drivers (DMA, Korean law, Dutch ruling, Epic ruling) into a clearly labeled 'compliance landscape as of <date>' section so the skill does not silently decay, or state that the compliance table must be re-verified before advising.
Convert the remaining descriptive passages into directives — e.g. add a short executable snippet for the apple-itunes-app banner meta tag and the Universal Link timeout fallback pattern, and phrase each compliance-table row as a 'do / don't' instruction rather than a fact.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is dense and expert-level — tables for drivers and compliance rules, a compact ASCII funnel, bullet mechanics — with zero explanation of concepts Claude already knows (no "what is Stripe", no deep-link tutorial). Not 5: the "Why Web-to-App Matters in 2025" section embeds time-sensitive, date-stamped regulatory claims ("Apple's EU DMA compliance + Korean law + Dutch dating-app ruling + DOJ Epic ruling") outside any 'deprecated/current' framing, and name-drops ("Used by Cal AI, Rise Sleep, Noom, Zing, Future") add padding; these could be trimmed or quarantined. Not 3: the padding is minor and localized, not a pattern of over-explanation. | 4 / 5 |
Actionability | For an instruction-only skill the guidance is highly concrete: exact quiz parameters ("6–12 questions", "30–60 seconds"), a placement rule ("Email capture after results revealed but before plan / price"), a named implementation detail ("<meta name=\"apple-itunes-app\">"), specific tool choices ("Branch Journeys", "AppsFlyer Smart Banner", "Funnelish, Heyflow, GetWaitlist"), and a fill-in Output Template that structures the deliverable. Not 5: a few sections stay descriptive rather than instructive (the drivers table and parts of the compliance matrix state facts but give no directive), and there is no executable snippet for the one code-adjacent piece (the banner meta tag / fallback timeout pattern is only named). | 4 / 5 |
Workflow Clarity | The sequence is clear and complete: Initial Assessment questions → pattern selection with explicit "Use when" criteria → the Pattern B mechanic as a step-by-step flow diagram → a five-step Output Template culminating in a compliance checklist and a measurement funnel. Not 5: checkpoints exist (the COMPLIANCE CHECKLIST with [ ] items, the MEASUREMENT funnel) but there is no validate-then-proceed gating or error-recovery loop — e.g. what to do if Universal Links verification fails is only implied by the Common Mistakes section. Not 3: the checklist and per-step conversion targets are explicit checkpoints, and this is not a destructive/batch operation requiring the cap. | 4 / 5 |
Progressive Disclosure | The file has no bundle at all (no references/, scripts/, or assets/), and at ~175 lines it carries several self-contained topics inline — the Apple/Google compliance matrix, quiz funnel mechanics, and smart-banner tooling — that would fit naturally as one-level-deep reference files for a skill family this clearly has siblings for (cross-skill handoffs name attribution-setup, onboarding-optimization, paywall-optimization). Not 2: the section structure is good with clear headers and navigable tables, so it is not "minimal structure". Not 4: the >50-line simple-skill exemption does not apply, and content that would be better split is inline with no external references at all. | 3 / 5 |
Total | 15 / 20 Passed |