Conciseness

The body contains no explanation of concepts Claude already knows (no preamble on what Fastly, caching, jq, or curl are) — every sentence carries Fastly-specific lore Claude cannot know, such as "a relative window opens and closes mid-bucket, so `--from "N days ago" --by day` returns N-1 buckets" and "`ts/h` on `rt.fastly.com` covers the last 120 seconds, not an hour". The example dates (2026-07-01T00:00:00Z) demonstrate the exact UTC-boundary format rule 2 depends on rather than encoding time-sensitive instructions, so they do not trigger the date penalty. This is the lean, competence-assuming 5 anchor, not the 4 anchor with trimmable over-explanation.

Actionability

Five complete, copy-paste-ready bash pipelines with jq cover the common cases (monthly hit ratio recomputed from counters, 5xx share correct on both service types, per-service bandwidth ranking driven from the service list, account usage totals, and a wall-clock-bounded live RPS sample), plus a command-selection table and a concrete curl example with the auth header. This matches the 5 anchor (fully executable, examples cover common cases) rather than the 4 anchor with minor gaps.

Workflow Clarity

The path from question to answer is unambiguous: pick the command from the table, apply the ten correctness rules, adapt a worked answer. Explicit validation checkpoints are present where they matter — "Read back `meta.from` / `meta.to`", "assert `meta` carries the filter you sent", "Check `fastly products -s ID` before concluding there is nothing to see" — and error recovery is delegated to references/debugging.md. Operations are read-only queries, so the destructive/batch validation cap does not apply, and under the simple-skill note the single action is unambiguous, fitting the 5 anchor over the 4.

Progressive Disclosure

The body keeps only the operating core (rules, command table, worked answers, traps) and moves exhaustive detail to three verified one-level-deep references, each clearly signaled with its content: "Endpoint paths, parameters and response shapes: [references/api.md]", "Field names and aggregation shape: [references/fields.md]", "Errors, empty data and wrong-scope symptoms: [references/debugging.md]". All three files exist (203/178/192 lines) with content matching their labels, and no reference nests further references, matching the 5 anchor. The single external pointer to the `fastly-cli` skill's flag matrix is clearly labeled but does not create buried navigation.

Specificity

The description enumerates ten concrete capabilities ("cache hit ratio, bandwidth, request counts, status-code and error rates, edge vs origin traffic, real-time requests-per-second, origin latency, per-domain traffic, account usage and billing totals") and names the exact tools it owns ("the `fastly stats` CLI commands and the Historical Stats, Real-Time and Origin/Domain Inspector HTTP APIs"). This matches the anchor for multiple specific concrete actions with comprehensive coverage; there is no generic filler, so it is not the level-4 anchor with minor gaps.

Completeness

It explicitly answers both questions: what it does ("Fastly traffic numbers..." plus the owned commands and APIs) and when to use it ("Use for any question that needs a number about how a Fastly service is performing or how much it is being used"). Both are concrete and explicit, matching the 5 anchor; the 4 anchor requires a 'when' that could be more specific, which is not the case here.

Trigger Term Quality

The terms are the natural phrases a user would say when needing this skill — "traffic numbers", "cache hit ratio", "bandwidth", "request counts", "error rates", "requests-per-second", "origin latency", "billing totals", "how much it is being used" — with synonyms covered (e.g. both "real-time requests-per-second" and performance/usage framings). Coverage is comprehensive for the domain rather than merely good with a few natural terms missing, so it fits the 5 anchor over the 4.

Distinctiveness Conflict Risk

It claims a clear niche (Fastly metrics) and explicitly delineates ownership ("Owns the `fastly stats` CLI commands and the Historical Stats, Real-Time and Origin/Domain Inspector HTTP APIs"), so it is unlikely to trigger for sibling Fastly configuration or logging skills. Distinct triggers and minimal conflict risk match the 5 anchor rather than the 4 anchor's minor overlap risk.