Saltar al contenido
TakoAPI
Volver a skills

sentiment-priority-scorer

por vishalgojha

Marketing & Sales
335Descargas según clawskills.sh
0Vistas
0 me gustaInicia sesión para dar me gusta a esta skill
—Aún sin valoraciones
Tu valoración

Resumen

Score normalized real-estate leads using sentiment, urgency, intent, recency, and record type to produce.

Descripción

Qué hace esta skill

Scores normalized real-estate leads across sentiment, urgency, intent, and recency dimensions to produce a deterministic 0-100 priority score and P1/P2/P3 bucket assignment. Operates read-only — no writes, messages, or outbound actions. Designed to sit in a multi-agent chain after lead normalization.

Deterministic weighted scoring with explicit evidence tokens makes priority decisions reproducible and auditable, unlike subjective or model-driven ranking that varies between runs.

Cuándo usarla

  • Ranking a callback queue before the morning sales shift
  • Triaging hundreds of inbound leads after a property launch campaign
  • Classifying buyer requirements vs inventory listings by follow-up urgency
  • Identifying P1 leads signaling immediate possession or same-day access
  • Sorting a stale lead backlog when team bandwidth is limited

Flujo de trabajo de ejemplo

ENTRADA

User asks: Rank these 50 normalized leads by callback priority

AGENTE
  1. 1

    Validates input against sentiment-priority-input.schema.json

  2. 2

    Computes sentiment_score, intent_score, recency_score, and urgency_score for each lead

  3. 3

    Applies record-type logic: boosts intent_score for buyer_requirement records, checks for high-action cues in listing text

  4. 4

    Calculates priority_score using the weighted formula and assigns P1, P2, or P3 bucket

  5. 5

    Returns scored records with plain-language evidence tokens explaining each score

SALIDA

Ranked list of leads with priority_score, bucket (P1/P2/P3), and evidence tokens per record

Instalación

clawhub install vishalgojha/sentiment-priority-scorer

Ejecuta el comando de la CLI o pega el prompt en Claude Code, Cursor, Codex o el agente de programación que uses.

Enlaces