Every pipeline decision is recorded with timestamp + SHA-256 hash. Immutable audit trail. Replay any hero's journey.
All recorded pipeline actions. Each writes to evidence_chain in DDB with {step, decision, detail, timestamp, hash}.
| Step | Records | Replay Command | Description |
|---|---|---|---|
le_backfill | COMPUTED | python3 -c "from intelligence_orchestrator import *" | Ligand efficiency computed from Vina + MW |
target_justification | PURSUE / HOLD | python3 -c "from target_justification import justify_target; print(justify_target('GENE'))" | Why pursue this target? ChEMBL + DepMap + pathway scoring |
ip_gate | DE_NOVO / REPURPOSE / CDX / DEPRIORITIZE | python3 -c "from ip_gate import run_ip_gate; print(run_ip_gate('GENE', hero_data))" | Unified IP classification: patent + ChEMBL phase + trials + expression |
ip_router | REPURPOSE_CONFIRMED / CDX_VALIDATED / PARKED | python3 -c "from repurpose_expression_matcher import check_expression_in_cancer; print(check_expression_in_cancer('GENE','colorectal'))" | Post-IP-gate routing: expression match, biomarker validate, or park |
tox_sel_backfill | COMPUTED | python3 -c "from rdkit.Chem import Descriptors; ..." | Toxicity (hERG risk) + selectivity (fsp3/promiscuity) from SMILES |
competitive_diff | PROCEED / PARK | python3 -c "from competitive_diff_scorer import compute_competitive_diff; print(compute_competitive_diff(hero))" | Is our compound differentiated vs trial competitors? |
agentic_retry | GPU_RESUBMIT | aws batch submit-job --job-name retry-GENE ... | Auto-retry with scaffold seeding + CRC enhancement |
evidence_assessment | authority=X.XXX | python3 -c "from evidence_gap_builder import assess_evidence; ..." | Multi-source evidence authority (inhibitors, structure, family, pathway) |
evidence_gap_builder | full_authority / evidence_missing | python3 -c "from evidence_gap_builder import build_gap_report; ..." | What evidence is missing? What assays to run? |
repurposing_candidate | Drug name | python3 -c "from repurpose_expression_matcher import check_expression_in_cancer; ..." | Approved drug found for this target |
chemistry_validation | PASS / FAIL | python3 -c "from chemistry_gate import validate_compound; ..." | PAINS, reactive groups, Lipinski, stability checks |
grade_assignment | A+ / A / B+ / B / C | python3 -c "from chemistry_gate import compute_grade; ..." | Final hero grade based on composite score |
gpu_dock | UPDATED | aws batch submit-job --job-name dock-GENE ... | GPU docking result written (Vina score, SMILES) |
brain_bus_sync | ALL_BRAINS_CONNECTED | python3 -c "..." | CortexQ brain signals synchronized |
cd /Users/georgesoto/projects/edice-r
python3 << 'EOF'
import sys, json; sys.path.insert(0, 'agents')
import os; os.environ['AWS_PROFILE'] = 'edice-r-admin'
import boto3
ddb = boto3.Session(profile_name='edice-r-admin', region_name='us-east-1').resource('dynamodb')
hero = ddb.Table('cortex-heroes').get_item(Key={'pk': 'GENE_NAME'}).get('Item', {})
chain = hero.get('evidence_chain', [])
print(f"Evidence chain for {hero.get('pk')} ({len(chain)} entries):")
for entry in chain:
e = json.loads(entry) if isinstance(entry, str) else entry
print(f" [{e.get('timestamp','')}] {e.get('step'):25s} → {e.get('decision'):20s} | {e.get('detail','')}")
EOF
# Replay IP Gate decision for a specific hero
python3 -c "
import sys; sys.path.insert(0, 'agents')
from ip_gate import run_ip_gate
result = run_ip_gate('KRAS_CRC', {
'chembl_max_phase': 4,
'clinical_trial_count': 795,
'cancer_type': 'colorectal'
})
print(result)
"
# Run nightly orchestrator locally (reads from DDB, writes evidence) cd /Users/georgesoto/projects/edice-r AWS_PROFILE=edice-r-admin python3 -c " import sys; sys.path.insert(0, 'agents') from intelligence_orchestrator import run_nightly_intelligence result = run_nightly_intelligence() import json; print(json.dumps(result, indent=2, default=str)) "
# Each entry has a hash: sha256(gene:step:decision:timestamp)
# Verify no entries were tampered with:
python3 -c "
import sys, json, hashlib; sys.path.insert(0, 'agents')
import boto3
ddb = boto3.Session(profile_name='edice-r-admin', region_name='us-east-1').resource('dynamodb')
hero = ddb.Table('cortex-heroes').get_item(Key={'pk': 'GENE_NAME'}).get('Item', {})
chain = hero.get('evidence_chain', [])
for entry in chain:
e = json.loads(entry) if isinstance(entry, str) else entry
expected = hashlib.sha256(f\"{hero['pk']}:{e['step']}:{e['decision']}:{e['timestamp']}\".encode()).hexdigest()[:16]
status = '✅' if e.get('hash') == expected else '❌ TAMPERED'
print(f\" {status} [{e['timestamp']}] {e['step']} → {e['decision']}\")
"
Loading hero list...