What

Adds rcl-score to the Data & Analysis section.

Skill summary

rcl-score turns a two-person chat history (WeChat CSV / JSON / TXT) into a relationship K-line candlestick chart — the trading-style visualization, applied to relational data.

• The IDE LLM scores every conversation turn with v3.1 semantic deltas across 10 dimensions (affection, engagement, care, conflict, tension, investment, awkwardness, future_orientation, vulnerability, shared_identity)
• A local Python framework converts message-level scores into OHLC bars across 11 timeframes (5m → 季 → 年)
• Local web frontend renders the chart with MA / Bollinger Bands / MACD / RSI / KDJ — same indicators stock traders use
• Per-bar attribution: hover any candle → see which dimension drove the move + the 4 most representative key messages from that period
• Everything runs locally. Chat data never leaves the user's machine. Scoring happens in the user's IDE LLM (Claude Code / Codex / etc.), not via external API.
• Bilingual: SKILL.md frontmatter EN, body 中文 + EN scoring rules

Files changed

• README.md — single line added under Data & Analysis, alphabetical placement between postgres and recursive-research

License

MIT — see LICENSE