Conciseness

The body is mostly lean — command tables, curl/jq snippets, and a diagnostics table with little concept explanation — but contains several redundant repetitions: the 'companies/all.json is large (~5,700 companies), prefer specific endpoints' warning appears in the Step 2 table, Key rule 3, and the diagnostics table, and batch/tag lowercase-hyphen naming is stated in three places (the two format sections and Key rules 5–6). That fits 'efficient; minor instances of over-explanation that could be trimmed' better than the lean top anchor, but well above the 'mostly efficient but includes some unnecessary explanation' anchor at 3.

Actionability

Every step is copy-paste executable: a concrete base URL, working curl|jq one-liners for fetching, counting, filtering, field extraction, and case-insensitive search, a table of ready-made jq filters, and exact meta.json commands for discovering valid names. Examples cover the common cases, matching the 'fully executable; copy-paste ready' top anchor.

Workflow Clarity

The five-step sequence (verify prerequisites → match request to endpoint → execute → present → diagnose) is clearly ordered, with an explicit prerequisite check that flags CURL_MISSING/JQ_MISSING, a diagnostics table mapping error → cause → fix, and meta.json-based recovery for invalid names — a genuine validate → fix → retry loop. This is read-only with no destructive or batch operations, so no validation cap applies, and it matches the 'clear sequence with explicit validation steps; feedback loops for error recovery' anchor.

Progressive Disclosure

SKILL.md is a genuine overview (request→endpoint table, general patterns, jq filters) that points clearly and in two places to a real single reference file, 'references/api_reference.md', for the full company schema and research workflows. The bundle confirms that structure: one references file, one level deep, no further nesting, and no scripts/assets — exactly the 'clear overview with well-signaled one-level-deep references' top anchor.

Specificity

The description lists many concrete capabilities — 'company profiles, batch rosters (e.g. W25, S24), companies by industry or tag, top companies, who is hiring, founder-diversity lists, and overall YC stats' — giving comprehensive, non-generic coverage of what the skill does. It names the exact data source (the public yc-oss API) rather than abstract language, matching the anchor for multiple specific concrete actions with comprehensive coverage; there is no vagueness or padding to drop it below 5.

Completeness

It explicitly answers both questions: 'what' is the enumerated lookup capabilities over the yc-oss dataset, and 'when' is an explicit 'Use this skill whenever the user asks about YC-backed startups, a YC batch, YC companies in a sector or tag or that are hiring...' with concrete trigger phrases. This matches the anchor that clearly and explicitly answers both what AND when; the score-4 anchor ('when' could be more explicit) does not apply since the trigger list is specific and multi-condition.

Trigger Term Quality

Natural user phrasings are covered comprehensively with synonyms: 'YC-backed startups', 'a YC batch', 'YC companies in a sector or tag or that are hiring', 'the Y Combinator portfolio', 'startup and venture research', plus short batch forms 'W25, S24'. These are all phrases a user would naturally say when needing this skill, and both 'YC' and 'Y Combinator' variants appear; no natural term is obviously missing, so it fits the top anchor rather than the 'a few natural terms missing' anchor at 4.

Distinctiveness Conflict Risk

It occupies a clear niche — read-only lookup against a specific named dataset (the yc-oss API) — with triggers all scoped to YC ('YC-backed startups', 'YC batch', 'Y Combinator portfolio'). The qualifier 'or startup and venture research that draws on YC data' even bounds the broadest trigger, minimizing conflict with general startup-research skills, so it fits the 'clear niche with distinct triggers' top anchor.