Conciseness

Mostly efficient with concrete, domain-specific content (ungrammar legend, naming conventions, real biome_parser APIs), but the 'Common Patterns' section restates patterns already shown in the workflow examples and Tips, and full lexer/token-source trait boilerplate could be trimmed. Matches the score-2 anchor ('Mostly efficient but includes some unnecessary explanation or could be tightened'); not score 3 because of redundant repetition, not score 1 because it avoids explaining general concepts Claude already knows.

Actionability

Provides fully executable Rust using real biome_parser APIs (parse_if_statement, ParseSeparatedList, ParseRecoveryTokenSet, StrictMode.excluding_syntax) plus concrete commands ('just install-tools', 'just gen-grammar html', 'cargo test') and specific file paths. Matches the score-3 anchor ('Fully executable code/commands; specific examples; copy-paste ready'); not score 2 because the code is real API usage rather than pseudocode, despite one placeholder (read_next_token) in the lexer example.

Workflow Clarity

Steps are clearly sequenced (Create Grammar -> Generate Parser -> Implement Lexer -> Token Source -> Parse Rules -> Lists -> Error Recovery -> Conditional Syntax -> Test Parser), but validation is end-loaded: only the final 'Test Parser' section with cargo test, with no mid-process checkpoints after codegen or after implementing parse rules. Matches the score-2 anchor ('Steps listed but validation gaps; sequence present but checkpoints missing or implicit'); not score 3 because of absent mid-process validation checkpoints, not score 1 because the sequence is explicit.

Progressive Disclosure

Well-organized into clear sections with a dedicated, one-level-deep '## References' list pointing to the Biome codebase (CONTRIBUTING.md, xtask/codegen/*.ungram, existing parsers). However, the body is a ~340-line monolithic document carrying multiple full code templates inline that could be split into reference files, fitting the score-2 anchor ('content that should be separate is inline'); not score 3 because content is not split across files, not score 1 because organization is clean with no nested references.

Specificity

Lists multiple concrete actions — 'implementing parsers with error recovery', 'adding parsing support for a new language', 'implementing error recovery in a parser', 'writing grammar definitions in .ungram format' — rather than naming only a domain. This matches the score-3 anchor ('Lists multiple specific concrete actions'); it is not score 2 because the actions are specific and several, not a single vague domain mention.

Completeness

Clearly answers both 'what' ('Guide for implementing parsers with error recovery for new languages in Biome') and 'when' with an explicit 'Use when...' trigger clause. This matches the score-3 anchor ('Clearly answers both what AND when with explicit triggers'); not score 2 because the 'when' is explicit, not merely implied.

Trigger Term Quality

The 'Use when...' clause covers natural trigger phrasings a Biome contributor would say ('adding parsing support', 'implementing error recovery', 'writing grammar definitions') plus six concrete language names (JavaScript, CSS, JSON, HTML, GraphQL) and '.ungram format'. Good coverage of natural domain terms, matching the score-3 anchor; not score 2 because common variations are well represented rather than missing.

Distinctiveness Conflict Risk

Narrow niche — parsers with error recovery for new languages in Biome, tied to .ungram format and named languages — making it unlikely to trigger for unrelated skills. Matches the score-3 anchor ('Clear niche with distinct triggers; unlikely to conflict'); not score 2 because the scope is sharply bounded rather than overlapping.