Conciseness

The body is dense and table-driven (Referral vs. Affiliate, Reward Types, Optimization Levers, Key Metrics) with no explanations of concepts Claude already knows, but minor padding remains — the persona preamble ("You know the difference between programs that compound and programs that collect dust") and a few narrative flourishes could be trimmed from the 275-line body.

Actionability

Concrete, numbers-grounded guidance throughout ("Reward should be ≥10% of first month's value", "cap at 30% of first payment", "Cookie window set (30 days minimum, 90 days for B2B)", named platforms Impact/ShareASale/Rewardful/PartnerStack, a worked tier example), but "Run `scripts/referral_roi_calculator.py`" references a script that does not exist in the bundle, leaving a small execution gap.

Workflow Clarity

All three modes give a clear numbered sequence, and advisory checkpoints are present ("Measure the impact before moving to the next lever", "10-affiliate pilot before scaling", "Confirm reward eligibility as soon as referral signs up"). Not 5 because the checkpoints are directional advice rather than explicit validate-fix-retry loops; not 3 because sequencing and checkpointing are clearly present and the domain is non-destructive, so no cap applies.

Progressive Disclosure

References are well-signaled and one level deep ("See [references/measurement-framework.md] for benchmarks by industry", "See [references/program-mechanics.md] for detailed program patterns"), but neither referenced file exists in the bundle — references/, scripts/, and assets/ directories are all absent — so the disclosure structure dead-ends. The ~275-line body also inlines content (affiliate launch checklist, reward-type tables) that the reference files were meant to carry.

Specificity

The description names the domain and several concrete actions ("design, launch, or optimize a referral or affiliate program") plus coverage areas ("program mechanics, incentive design, and optimization"), with only minor gaps — it stops short of enumerating the distinct sub-capabilities (incentive sizing, affiliate toolkit, measurement framework) that the body actually covers.

Completeness

Explicitly answers both halves: what it does ("design, launch, or optimize a referral or affiliate program... Covers program mechanics, incentive design, and optimization") and when to use it ("Use when they mention [concrete trigger phrases]"), matching the top anchor's what+when pattern with concrete triggers.

Trigger Term Quality

Comprehensive natural trigger terms with synonyms: 'referral program,' 'affiliate program,' 'word of mouth,' 'refer a friend,' 'incentive program,' 'customer referrals,' 'brand ambassador,' 'partner program,' 'referral link,' and 'growth through referrals' — the phrases a user would actually say when they need this skill.

Distinctiveness Conflict Risk

The referral/affiliate niche is clear and mostly distinct, but broad trigger terms like 'incentive program,' 'partner program,' and 'word of mouth' carry minor overlap risk with general marketing or go-to-market skills — noticeable above the midpoint but not the fully clean niche of a 5.