Checking the session, access policy, and current workspace data.
Fake Token + Compromised Admin + Oracle Manipulation
On April 1, 2026, Drift Protocol suffered a $286M triple-vector attack: a fake collateral token accepted by the vault, a compromised admin key that disabled circuit breakers, and manipulated oracle prices used by the perpetuals engine.
Imported-record overview
These fields reproduce or normalize claims in an imported research record. “Reported” describes provenance; it does not mean ChainRadar has confirmed the claim.
Treat this as a reported explanation to test against code, traces, transaction data, and other primary evidence.
Date-free learning path
Synthetic study sequence for learning purposes; timestamps and ordering may not reproduce the full verified incident chronology. Synthetic dates are deliberately omitted below.
STUDY PHASE
STUDY PHASE
ClearingHouse.initialize_spot_market(fakeUSDC_mint, decimals=6, symbol='USDC')STUDY PHASE
ClearingHouse.update_oracle_guard_rails(max_age=u64::MAX, confidence_band_pct=100)STUDY PHASE
Serum.placeOrder(side=BID, size=huge, price=$1800)STUDY PHASE
ClearingHouse.deposit(fakeUSDC, 10_000_000_000)ClearingHouse.open_position(SOL, 20x_long, $1800)STUDY PHASE
ClearingHouse.settle_pnl(position_key)Teaching model
Educational reconstruction intended to explain the reported attack pattern, not a transaction-level evidentiary trace.
Start at Stage 1 and follow the numbered arrows. Every label is derived from the imported flow actions and structure.
Values describe reported movement between stages. Repeated amounts are not additional losses.
| Step | From | To | Reported action | Value |
|---|---|---|---|---|
| 01 | Reported fake input | Drift Protocol system | Fake Collateral | $10B |
| 02 | Drift Protocol system | Settle At Fake Price | Settle At Fake Price | $285M |
| 03 | Settle At Fake Price | Value destination | Drained | $285M |
Imported indicators
Contract entries are reproduced from the local research archive. Confirm chain, bytecode, ownership, and incident relevance against primary evidence before using them in an investigation.
dRiftyHA39MWEi3m9aunc5MzRF1JYuBsbn6VPcn33UHDefensive prompts
These imported mitigation labels are study prompts, not a complete or validated remediation plan. Test them against the protocol's actual architecture and failure mode.
Investigator workflow
Turn this background study into questions for a separate, evidence-led investigation. Never promote an educational edge or imported label into a finding without corroboration.
Preserve official disclosures, on-chain transactions, code, retrieval times, and other evidence supporting each material claim.
Start from corroborated addresses or transactions and trace each hop without assuming the educational flow is exact.
Compare decoded calls, logs, bytecode, and state changes with the reported cause.
Record provenance for every claim and identify gaps that still require confirmation.
Source status: cataloged. Classification: educational. Risk methodology: heuristic only. Primary-source verification remains required for every material claim.