Go to file

Claude e05cd392cc Add batch blind-eval harness and style retrieval, verified end-to-end blind_eval.py runs PoC #1's blind-eval methodology over N held-out corpus dialogues automatically (generate_draft.py refactored to expose draft_reply() so both share the same drafting logic). retrieve_style.py implements the keyword/recency search from tech-design.md §2-1 and wires into generate_draft.py as --history, replacing hand-curated --style files. Bash access was intermittently restricted for part of this session (auto-mode safety classifier), so these were initially written and committed-pending without live execution. Now verified for real: generate_draft.py's existing behavior still holds after the draft_reply() refactor, blind_eval.py runs cleanly against val.jsonl, and retrieve_style.py's original weighted-sum scoring had a real bug -- recency drowned out keyword overlap for short Korean messages (particle attachment means "핀란드" and "핀란드는" don't share a token), so it was effectively returning the most recent messages regardless of topic. Fixed by ranking on (overlap, recency) instead of a weighted sum, confirmed the Finland-related exemplar now ranks first for a matching query.		2026-07-29 09:20:41 +00:00
docs	Add batch blind-eval harness and style retrieval, verified end-to-end	2026-07-29 09:20:41 +00:00
poc/tone-corpus	Add batch blind-eval harness and style retrieval, verified end-to-end	2026-07-29 09:20:41 +00:00
.env.example	Add root .env template for GEMINI_API_KEY	2026-07-29 08:37:29 +00:00
.gitignore	Add rule-based escalation gate before any LLM call	2026-07-29 08:56:12 +00:00
AGENTS.md	Sync planning docs with tentative Q1-Q7 decisions	2026-07-29 07:14:23 +00:00
CLAUDE.md	Add AGENTS.md and CLAUDE.md for Cursor/Claude project rules	2026-07-29 07:12:25 +00:00
README.md	Add one-page meeting review summary and wire up README	2026-07-29 07:39:06 +00:00

README.md

hikikomori

카카오톡 대안 메신저 아이디어 — "나를 대신해 남과 대화하는 AI 분신"이 있는 메신저 프로젝트.

문서

원본 회의 자료: docs/idea-meeting-2026-06-29.html
기획 진행 방법: docs/PLANNING.md
결정 로그 (Q1~Q7): docs/decision-log.md
Vision Doc: docs/vision.md
PRD (v1 클로즈드 베타): docs/PRD.md
기술 설계서: docs/tech-design.md
리스크 로그: docs/risk-log.md
로드맵: docs/roadmap.md
PoC #1·#3 실행 계획: docs/poc-plan.md / 실행 준비물: docs/poc-materials.md
Q3 자율성 수용성 인터뷰 가이드: docs/user-interview-guide.md
회의 리뷰 1페이지 요약: docs/meeting-review-summary.md

AI 에이전트 규칙

Cursor / 공통: AGENTS.md
Claude Code 진입점: CLAUDE.md

현재 단계

아이디어 발산 → 컨셉 심화 → 기획 문서 세트 작성 → PoC 실행 계획·준비물·클릭 프로토타입 완료. docs/decision-log.md의 결정은 아직 잠정안이므로, 실제 회의에서 확정/수정이 필요합니다. 다음 단계는 PoC #1·#3과 Q3 인터뷰를 실제로 진행하는 것입니다 — 계획과 자극재는 이미 준비돼 있으니 참가자 모집부터 시작하면 됩니다 (docs/meeting-review-summary.md 참고).