resume-initlisted

# resume-init ## Overview Three-round structured interview that builds a comprehensive personal profile stored in `state.json`. Each round is an independent context window that loads its own reference file plus the current `state.json`. `state.json` is the **only** data bridge between rounds — all collected data is written there immediately after each answer, so interruptions never lose progress. ## Zero-Interview Quick Path When no `state.json` exists and user runs with `--quick`: 1. Auto-detect `git_authors` from local git log 2. Create minimal `state.json` with detected authors and placeholder profile 3. Skip directly to `resume-scan` + `resume-generate` 4. User can always run `wei-resume init` later to enrich the profile ## Interview Rounds ### Round 1: Deterministic Information (3-5 min) - Name, contact, links, education, career overview, git authors - Load reference: `references/round1-basic.md` - Output: `state.json` sections `profile.basic`, `.education`, `.career`, `.git_authors` ### Round 2: Exploratory Mining (5-8 min) - Tech highlights, team role, proud projects, certs/OSS/articles, target direction - Load reference: `references/round2-explore.md` - Output: `state.json` section `profile.strengths`, `.target_direction` ### Round 3: Deep Interview (12-18 min) - Merged personality + AI cognition, adaptive topic pools - Load reference: `references/round3-deep.md` - Output: `state.json` sections `profile.persona`, `.ai` ### Import Round (optional, 5-10 min) -