Conciseness

The body is lean: it contains no explanation of what NFC is, how delegates work, or other concepts Claude already knows — nearly every line is executable code, exact entitlement/Info.plist keys, or non-obvious framework facts (e.g., first-URI-record routing, iPhone 7/XS hardware cutoffs). It is not 4 because the only duplication (the 2-sentence intro) doubles as a body summary and there is no padding to trim.

Actionability

Guidance is fully executable: complete, copy-paste Swift for availability checking, an NDEF reader delegate, tag connection with status branching, writing with validation, payload construction, and background-tag handling via NSUserActivity, plus exact keys like com.apple.developer.nfc.readersession.formats with the current TAG value. It is not 4 because the common cases (read, write, background) are each covered with complete runnable code and concrete error-code filtering.

Workflow Clarity

Multi-step processes are clearly sequenced with validation checkpoints: numbered project-configuration steps, NFCNDEFReaderSession.readingAvailable checked before creating sessions, queryNDEFStatus gating reads/writes, error filtering that distinguishes user cancellation, and a 15-item review checklist with feedback loops (invalidate with errorMessage on failure). It is not 4 because validation is explicit rather than implied, and the semi-destructive write path is guarded by status and capacity checks, so no cap applies.

Progressive Disclosure

The body is a well-organized overview that keeps advanced material (APDU commands, MIFARE/FeliCa operations, NDEF locking, SwiftUI scanner) in a single real one-level-deep reference, clearly signaled at point of use ("Load nfc-patterns.md for protocol-specific connection, APDU, command, and response handling") and with an anchored link into its Parsing NDEF Payload Content section; references/nfc-patterns.md exists and contains that section. It is not 4 because structure and navigation are exactly the clear-overview-with-well-signaled-references pattern.

Specificity

The description enumerates multiple concrete actions — "Read and write NFC tags", "scanning NDEF tags", "reading ISO7816/ISO15693/FeliCa/MIFARE tags", "writing NDEF messages", "handling NFC session lifecycle", "configuring NFC entitlements", "implementing background tag reading" — with comprehensive coverage of the framework's surface, matching the anchor-5 example. It is not 4 because there are no minor gaps in coverage; every major CoreNFC capability is named.

Completeness

It explicitly answers both questions: what ("Read and write NFC tags using CoreNFC") and when ("Use when scanning NDEF tags, ... or implementing background tag reading in iOS apps") with concrete trigger phrases. It is not 4 because the when-clause is fully explicit and enumerates six concrete triggers rather than a generic condition.

Trigger Term Quality

Natural user phrases are comprehensively covered: "NFC tags", "scan", "NDEF", the four tag-family names, "entitlements", and "background tag reading" — the terms a user would naturally say when needing this skill. It is not 4 because no common synonym or variation (e.g., "read a tag", "write NFC", "NFC reader") is missing.

Distinctiveness Conflict Risk

The description occupies a clear niche (CoreNFC on iOS) with distinct triggers — no other plausible skill covers NFC tag reading/writing, so conflict risk is minimal. It is not 4 because there is no closely related skill domain with which these triggers could overlap.