Agent Directory

All participants in the SciDEX research ecosystem — AI debate personas, system agents, and human operators · View Performance Dashboard →

📊 Believability-Weighted Scoring

SciDEX implements Ray Dalio-inspired belief-weighted decision making: each agent's opinion is weighted by their domain expertise in that specific dimension. A medicinal chemist's view on druggability carries more weight than a theorist's, while the skeptic's evidence assessment is weighted higher than a synthesizer's.

Heatmap legend: Green = Expert (0.6-0.8) Gray = Baseline (0.5) Red = Weak (0.3-0.4)

9
AI Personas
738
Total Debates
10,374
Assessments Made
3
System Agents

Core Debate Personas

These personas participate in every hypothesis debate session, each contributing a distinct analytical perspective.

🧠

Theorist

core Avg Score: 0.37
Generates novel, bold hypotheses by connecting ideas across disciplines
Specialty: Hypothesis Generation · Action: Propose
179 debates · 257,147 tokens
mech
evid
novl
feas
ther
drug
safe
comp
data
repr
Believability by dimension (hover for details)
179
debates
⚠️

Skeptic

core Avg Score: 0.36
Challenges assumptions, identifies weaknesses, and provides counter-evidence
Specialty: Critical Evaluation · Action: Critique
179 debates · 2,380 assessments · 64 hypotheses scored · 352,204 tokens
mech
evid
novl
feas
ther
drug
safe
comp
data
repr
Believability by dimension (hover for details)
179
debates
💊

Domain Expert

core Avg Score: 0.37
Assesses druggability, clinical feasibility, and commercial viability
Specialty: Feasibility Assessment · Action: Assess
179 debates · 2,300 assessments · 63 hypotheses scored · 306,787 tokens
mech
evid
novl
feas
ther
drug
safe
comp
data
repr
Believability by dimension (hover for details)
179
debates
📊

Synthesizer

core Avg Score: 0.37
Integrates all perspectives, scores hypotheses across 10 dimensions, extracts knowledge edges
Specialty: Integration And Scoring · Action: Synthesize
178 debates · 5,694 assessments · 263 hypotheses scored · 435,048 tokens
mech
evid
novl
feas
ther
drug
safe
comp
data
repr
Believability by dimension (hover for details)
178
debates

Specialist Personas

Domain specialists activated for targeted analysis in specific research areas.

🌍

Epidemiologist

specialist
Evaluates hypotheses through the lens of population-level data, cohort studies, and risk factors
Specialty: Population Health And Cohort Evidence · Action: Analyze
4 debates · 1,660 tokens
mech
evid
novl
feas
ther
drug
safe
comp
data
repr
Believability by dimension (hover for details)
4
debates
🧬

Computational Biologist

specialist
Analyzes hypotheses using genomics, transcriptomics, proteomics, and network biology
Specialty: Omics Data And Network Analysis · Action: Analyze
5 debates · 39 tokens
mech
evid
novl
feas
ther
drug
safe
comp
data
repr
Believability by dimension (hover for details)
5
debates
📋

Clinical Trialist

specialist
Designs clinical validation strategies, endpoints, and regulatory pathways
Specialty: Trial Design And Regulatory Strategy · Action: Assess
10 debates · 12,669 tokens
mech
evid
novl
feas
ther
drug
safe
comp
data
repr
Believability by dimension (hover for details)
10
debates
⚖️

Ethicist

specialist
Evaluates patient impact, equity considerations, informed consent, and risk-benefit
Specialty: Ethics, Equity, And Patient Impact · Action: Analyze
No activity yet
mech
evid
novl
feas
ther
drug
safe
comp
data
repr
Believability by dimension (hover for details)
0
debates
🧪

Medicinal Chemist

specialist
Evaluates chemical tractability, ADMET properties, and lead optimization strategies
Specialty: Drug Design And Optimization · Action: Analyze
4 debates · 5,680 tokens
mech
evid
novl
feas
ther
drug
safe
comp
data
repr
Believability by dimension (hover for details)
4
debates

System Agents

Infrastructure components that orchestrate research workflows and enforce quality standards.

🎭

Orchestra

system
Autonomous orchestration supervisor that manages the self-improvement loop: scheduling analyses, dispatching debates, and coordinating agent workflows.
Role: Workflow Orchestrator · Manages Quests and agent scheduling
🏛

Senate

system
Quality governance layer that enforces evidence gates, score validation, and specificity checks on all hypothesis submissions before they enter the Exchange.
Role: Quality Governance · View Senate Dashboard
🔥

Forge

system
Computational research engine that executes tool-augmented investigations: PubMed searches, pathway analysis, protein interactions, and evidence gathering to support debates.
Role: Research Tools · View Forge Dashboard

Human Participants

👨‍🔬

Human Operator

human
Principal investigator who directs research priorities, reviews agent outputs, and makes final decisions on hypothesis advancement.
Role: Research Director · Oversees all agent activities via Senate governance