Follow the authoring brief at /Users/evin/natively-cluely-ai-assistant/benchmarks/natively-answer-quality/dataset/AUTHORING-v2.md (and the v1 brief it points to). Do NOT read any other file under benchmarks/natively-answer-quality/dataset/ except categories.json and pi/profiles.json (if your mode needs PI). This is a HELD-OUT set: it must be written independently.

ADDENDUM — mode `looking-for-work`, partition `holdout`.
- File 1: /Users/evin/natively-cluely-ai-assistant/benchmarks/natively-answer-quality/dataset/holdout/looking-for-work.json with "partition": "holdout". EXACTLY 30 items, at least 3 chains of 2-3 turns. Context ids prefixed "HJOB-CTX-", local_ids "HJOB-001".. . Write your own new contexts: 1 grounded reference doc (e.g. a portfolio/project write-up or company research notes), 1 irrelevant, 1 conflicting. Speaker mix: about 80% other (interviewer), 20% user (candidate typing). Use PI per the brief (read pi/profiles.json): at least 20 of 30 items carry a pi_ref, covering null/resume-only/JD-only/both/both+doc conditions and at least 1 profile switch (PI-B after PI-A). Categories: "looking-for-work" in /Users/evin/natively-cluely-ai-assistant/benchmarks/natively-answer-quality/dataset/categories.json.
- File 2: /Users/evin/natively-cluely-ai-assistant/benchmarks/natively-answer-quality/dataset/final/looking-for-work-hard.json with "partition": "final-hard". EXACTLY 10 items, all HARD grounding tests: every item either depends on an attached doc / prior_transcript, or the PI profile (needles + evidence_required true), or is a trap where the needed fact is deliberately absent from the attached material (claim_trap set, needles []). Context ids "FJOB-CTX-", local_ids "FJOB-001".. ; write 1-2 new grounded docs and 1 conflicting doc for this file. No chains needed (at most 1).
