Conciseness

The ~550-line body inlines seven full Python classes plus a 'Core Concepts' section explaining things Claude already knows ("Monthly (most common for SaaS)", a subscription state diagram, what proration is), and it duplicates material (dunning retry logic, invoice construction appears in two places). This matches anchor 2 ('noticeably verbose; several unnecessary explanations or padded sections'); it avoids 1 only because the domain-specific implementation details (retry schedules, proration formulas, tiered pricing) carry some non-obvious value.

Actionability

The code is concrete and logically complete for the domain (proration math, tiered pricing, dunning schedules) but not executable: the Quick Start imports a fictional `billing` module, `stripe` is used without import, helpers like `generate_id`, `get_subscription`, `UsageRecord`, `send_email`, and `calculate_volume_pricing` are undefined, and `validate_vat_number`/`to_pdf` are `pass` stubs. This sits squarely on anchor 3 ('some concrete guidance but incomplete... missing key details') — above 2 because real, specific logic is present, below 4 because nothing runs as written.

Workflow Clarity

A sequence exists in `process_billing_cycle` (check due → generate invoice → charge → mark paid or start dunning) and the dunning retry flow is ordered, but there are no validation checkpoints, no verify/retry loops for the charge-and-cancel operations, and no guidance for sequencing an actual implementation. Billing is a batch, financially destructive domain, so per the rubric's cap, missing validation holds this at anchor 3 ('sequence present but checkpoints missing or implicit').

Progressive Disclosure

The Resources section lists 8 bundle files (`references/billing-cycles.md` through `assets/dunning-policy.yaml`), but none of these files exist in the bundle — every referenced path is dangling — while the ~450 lines of implementation code those files should hold are inlined in SKILL.md itself. This matches anchor 2 ('content that clearly belongs in separate files is inlined'); it does not reach 3 because the references, though clearly signaled, are broken, and the body is effectively the monolithic pattern the rubric penalizes.

Specificity

"Build automated billing systems for recurring payments, invoicing, subscription lifecycle, and dunning management" lists four concrete capability areas ("recurring payments", "invoicing", "subscription lifecycle", "dunning management"), matching the 'several specific actions; minor gaps' anchor. It falls short of 5 because the anchor-5 level would cover the domain comprehensively — proration, tax calculation, and usage-based billing (all core parts of the skill body) are omitted.

Completeness

Both parts are present: a clear 'what' ("Build automated billing systems for recurring payments, invoicing, subscription lifecycle, and dunning management") and an explicit 'when' ("Use when implementing subscription billing, automating invoicing, or managing recurring payment systems"). It stays at 4 rather than 5 because the 'when' clause largely restates the 'what' instead of adding concrete, differentiated trigger phrases (e.g., 'when the user mentions failed payments, plan upgrades, or invoice generation').

Trigger Term Quality

The trigger clause includes natural phrases users would say: "implementing subscription billing", "automating invoicing", "managing recurring payment systems" — good keyword coverage matching anchor 4. Not 5 because common variations are missing ("proration", "failed payments", "past due", "billing cycles", "tax/VAT"), and "dunning" appears only as jargon without a plain-language synonym.

Distinctiveness Conflict Risk

"Billing automation" with triggers like "subscription billing" and "dunning management" occupies a clear niche with distinct triggers, matching anchor 4. Not 5 because there is minor overlap risk with closely related skills a user might invoke instead, such as generic payment-processing or Stripe-integration skills.