lovstudio

Do They Love Me Skill — Communication skill for Claude Code

Communication community

Turn a two-person WeChat chat export into an evidence-based relationship analysis: interaction metrics, local work/affection/life semantic layering, disclosed error bounds, and an auditable mobile inf.

How to install Do They Love Me Skill

This entry records only its repository, not the path inside it, so there is no exact command to give. Open lovstudio/do-they-love-me-skill and copy the folder into ~/.claude/skills/, or the file into ~/.claude/agents/.

What Do They Love Me Skill does

Alternatives in Communication

Pregunta Cualquier Cosa Sobre El Código — /understand-chat How does the payment flow work 8.2k ★
Pestle Analysis — PESTLE analysis across Political, Economic, Social, Tech, Legal, Environmental 7.8k ★
YouTube Clipper Skill — Download videos, generate semantic chapters, clip segments, translate subtitles to bilingual format, and burn 1.6k ★

README

lov-do-they-love-me

![Version](https://img.shields.io/badge/version-0.2.0-CC785C)

把两个人的微信私聊做成有证据的恋爱指数分析：量化互动节奏，用本地模型把消息分成工作／情感／生活三类，最后产出一张手机竖版信息图。结论必须带口径与误差， **工作内容不计入情感热度**。

安装

npx skills add lov-do-they-love-me -g -y

从源码安装（维护者）

在本仓库根目录执行：

export SKILL_SOURCE_DIR="$(pwd)"
mkdir -p ~/.agents/skills
ln -sfn "$SKILL_SOURCE_DIR" ~/.agents/skills/lov-do-they-love-me
for host in ~/.claude/skills ~/.codex/skills; do
  mkdir -p "$host"
  ln -sfn ../../.agents/skills/lov-do-they-love-me "$host/lov-do-they-love-me"
done
readlink -f ~/.claude/skills/lov-do-they-love-me   # 应解析到本目录

用户 Profile（跨 session）

每个生成的 Skill 都会在 `skill.yaml` 中声明 `user-profile/v1`，并从共享 Profile 读取用户、品牌、工作区和本 Skill 的长期记录。用户直接说出的持久偏好或品牌事实由 `scripts/profile_store.py` 写回 Profile；源代码保持可移植。

详见 [`references/user-profile.md`](references/user-profile.md)。

使用

export SKILL_DIR="$SKILL_SOURCE_DIR"

# 1) 取数（交给 lov-wdb-cli，隔离副本只读）
python3 "$WDB_SKILL/skills/wdb-query/scripts/wdb_cli.py" chats \
  --contact "<昵称>" --from 2026-07-18 --to 2026-09-11 --limit 5000 --format jsonl > messages.jsonl

# 2) 量化 + 日历矩阵
python3 "$SKILL_DIR/scripts/chat_metrics.py"    --messages messages.jsonl --me <我> --other <对方> --out metrics.json
python3 "$SKILL_DIR/scripts/matrix_dataset.py"  --messages messages.jsonl --me <我> --other <对方> --out matrix.json

# 3) 语义分层（规则层 + 本地模型）+ 校准
python3 "$SKILL_DIR/scripts/export_text.py"     --messages messages.jsonl --me <我> --other <对方> --out label_input.jsonl
# 先校准：示例必须在校准集之外，否则算出来的是假准确率
python3 "$SKILL_DIR/scripts/semantic_label.py"  --input label_input.jsonl --out gold_pred.jsonl \
  --eval-gold gold.json --report accuracy.json
# 再全量标注（--context 40 会带上一条上文，短句才判得准）
python3 "$SKILL_DIR/scripts/semantic_label.py"  --input label_input.jsonl --out labels.jsonl \
  --model qwen3:8b --batch 20 --context 40
python3 "$SKILL_DIR/scripts/topic_composition.py" --labels labels.jsonl --out composition.json --accuracy accuracy.json

# 4) 出图，再交给 lov-mobile-i