The body is lean and code-forward, assuming Claude's competence — e.g. "graphql() from graphql-js executes against the schema directly. No server. No HTTP. No middleware." It avoids explaining concepts Claude already knows (no "what is GraphQL" preamble) and every section earns its tokens. Not score 2 because there is no padded or unnecessary explanation to tighten.

Provides fully executable, copy-paste-ready TypeScript (SchemaBuilder setup, graphql() unit test, printSchema snapshot, Apollo/Yoga wiring, relay plugin, context-value tests) plus concrete commands (npm install --save-dev, npm test -- --updateSnapshot, CI yaml). Matches the score-3 anchor and exceeds the pseudocode/incomplete score-2 anchor.

A clear topical progression (Install → Build schema → Unit-test resolver → Snapshot → Integration → Running → CI) with explicit validation feedback loops: the snapshot "fails and reviewers see the diff" and CI "--ci flag prevents --updateSnapshot and fails on mismatch", plus an anti-patterns table acting as a checklist. Not score 2 because validation checkpoints and error-recovery guidance are present rather than missing.

Well-organized into clear sections with a signaled References list, but the entire skill is a single inline file with no one-level-deep reference files (references/scripts/assets do not exist) and advanced material (plugin-contract and context-required tests) that could be split is inline. This fits the score-2 anchor (some structure, content that should be separate is inline) rather than the score-3 anchor requiring well-signaled separate reference files.

Lists multiple concrete actions with named primitives — "testing the SchemaBuilder output (lexicographicSortSchema + printSchema for snapshot tests)", "testing resolvers via the standard graphql() function", "integration with Apollo Server / GraphQL Yoga", and "code-first builder unit tests". This matches the score-3 anchor (multiple specific concrete actions) and exceeds the score-2 anchor, which only names a domain and some actions without this breadth.

Explicitly answers both what (the four testing patterns plus the SchemaBuilder API surface) and when via an explicit "Use when testing a Pothos-built schema before or alongside the server-runtime tests" clause. This is the score-3 anchor (clearly both what AND when with explicit triggers), not score 2 where when is only implied.

Covers natural terms a Pothos user would actually say — "Pothos", "GraphQL", "schema-builder", "snapshot tests", "resolvers", "Apollo Server", "GraphQL Yoga", "testing" — and the explicit trigger "Use when testing a Pothos-built schema". Not score 2 because common variations are present rather than merely "some relevant keywords".

Occupies a clear niche (Pothos code-first schema-builder testing) with distinct triggers and explicitly distinguishes sibling runtimes (apollo-server-tests, graphql-yoga-tests) as separate skills. Unlikely to trigger for the wrong skill, matching the score-3 anchor rather than the overlapping score-2 anchor.