jmanhype

Jtbd Idea Validator Agent — Data skill for Claude Code

Data community

AI-powered Jobs-to-be-Done Idea Validator Agent with flexible schema and enhanced analysis pipeline.

How to install Jtbd Idea Validator Agent

This entry records only its repository, not the path inside it, so there is no exact command to give. Open jmanhype/jtbd-idea-validator-agent and copy the folder into ~/.claude/skills/, or the file into ~/.claude/agents/.

What Jtbd Idea Validator Agent does

AI-powered Jobs-to-be-Done Idea Validator Agent with flexible schema and enhanced analysis pipeline.

Alternatives in Data

Notion Sync — /notion-sync - Push Ranked Jobs and Applications to a Notion Database 36.6k ★
User Segmentation — Segment users by behavior, JTBD, and needs from feedback data 7.8k ★
Debug Plus — Enhanced debug with swarm integration and prevention pipeline 508 ★

README

jtbd-idea-validator-agent

A DSPy-based agent that validates business ideas using Jobs-to-be-Done (JTBD) methodology. Takes a structured idea description, runs it through 4 analysis stages, and produces scored reports with charts and exportable data.

What it does

Input: a JSON file describing a business idea (problem, solution, target customer, hunches). Output: scored analysis with assumption deconstruction, JTBD job statements, competitive moat assessment, and actionable validation plans.

The 4 pipeline stages:

Stage	DSPy Module	Output
Deconstruct	`DeconstructSig`	Assumptions with confidence levels (1-3)
Jobs	`JobsSig`	5 JTBD statements, each with Four Forces (push/pull/anxiety/inertia)
Moat	`MoatSig`	Innovation layer analysis (Doblin framework)
Judge	`JudgeScoreSig`	Scores across 5 criteria, 0-10 scale

Scoring criteria: Underserved Opportunity, Strategic Impact, Market Scale, Solution Differentiability, Business Model Innovation.

Dual-judge arbitration

By default, 2 independent judge passes run and scores are merged with tie-breaking logic. Disable with `JTBD_DOUBLE_JUDGE=0`.

Quick start

pip install -e .

export OPENAI_API_KEY=...  # or ANTHROPIC_API_KEY

python run_direct.py examples/rehab_exercise_tracking_rich.json

Output goes to a timestamped directory:

reports//
  gamma/presentation.md    # Gamma-ready markdown
  gamma/presentation.html  # Preview
  csv/                     # Structured exports
  json/analysis_data.json  # Raw data
  assets/                  # Radar, waterfall, Four Forces charts

Configuration

Variable	Default	Purpose
`JTBD_DSPY_MODEL`	`gpt-4o-mini`	LLM model
`JTBD_LLM_TEMPERATURE`	`0.2`	Response randomness
`JTBD_DOUBLE_JUDGE`	`1`	Dual-judge arbitration
`JTBD_JUDGE_COMPILED`	(none)	Path to GEPA-compiled judge
`JTBD_LLM_SEED`	`42`	Repro