Score products on 10 evidence/regulatory dimensions and compute derived risk flags
Apply a standardized scoring model to biotech products based on evidence quality and regulatory setup, not stock price.
| Dimension | What it measures |
|---|---|
| Evidence Maturity | How far the clinical evidence has progressed |
| Endpoint Clarity | Is the primary endpoint well-defined and accepted by FDA? |
| Trial Design Quality | Randomized, controlled, adequate sample size? |
| Regulatory Advantage | Does it have designations (BTD, Fast Track, Orphan, Priority)? |
| Unmet Need Severity | How severe is the disease with how few alternatives? |
| Mechanism Plausibility | Is the MOA supported by biological rationale and prior data? |
| Manufacturing Complexity Risk |
| CMC risk (cell therapy > small molecule) |
| Safety Uncertainty | Known safety signals or class effects? |
| Sponsor Disclosure Quality | Transparent data sharing, consistent messaging? |
| Near-term Catalyst Density | How many meaningful milestones in the next 6 months? |
| Flag | Logic |
|---|---|
binary_event_risk | High if single catalyst determines product fate |
regulatory_visibility | High if FDA has official action date or accepted submission |
science_readthrough_value | High if mechanism has implications beyond lead indication |
crowded_indication_penalty | High if >5 competitors in same indication/phase |
approval_path_complexity | High if novel endpoint, first-in-class, or combo requirements |
scripts/score_product.pyInteractive or batch scoring. Accepts product data and returns composite scores + derived flags.
# Score a single product interactively
python .agent/skills/product_scoring/scripts/score_product.py --product-id PROD_001
# Batch score all active products
python .agent/skills/product_scoring/scripts/score_product.py --batch --active-only