While auditing slot-based interest accrual across Solana lending forks, I measured your deployment's realized borrow rates against the configured rates. The program accrues per slot with SLOTS_PER_YEAR = 63,072,000 (a 500ms slot assumption); mainnet slots have averaged ~396ms over the last 8 months and ~318ms over the last week. Every reserve realizes 1.26x-1.57x its configured APR - a 10% APR loan currently charges ~13.3%. Total excess across active reserves: ~$6.3M/yr at current state (table in the report).
Repro (zero deps, offline, ~2s): https://github.com/mv-reyes/solend-rate-drift-audit - npm test. The scanner re-verifies the sha256-pinned capture (slot 444,290,010) before parsing, so you can check the evidence before trusting my math. Full writeup in REPORT.md, including rate-curve validation gaps and the unchecked extra-oracle read.
While auditing slot-based interest accrual across Solana lending forks, I measured your deployment's realized borrow rates against the configured rates. The program accrues per slot with SLOTS_PER_YEAR = 63,072,000 (a 500ms slot assumption); mainnet slots have averaged ~396ms over the last 8 months and ~318ms over the last week. Every reserve realizes 1.26x-1.57x its configured APR - a 10% APR loan currently charges ~13.3%. Total excess across active reserves: ~$6.3M/yr at current state (table in the report).
Repro (zero deps, offline, ~2s): https://github.com/mv-reyes/solend-rate-drift-audit - npm test. The scanner re-verifies the sha256-pinned capture (slot 444,290,010) before parsing, so you can check the evidence before trusting my math. Full writeup in REPORT.md, including rate-curve validation gaps and the unchecked extra-oracle read.