Security review — 1 potential issues found

Security concerns in zama-ai/fhevm

I was looking through the codebase and found a few patterns that might be security-sensitive. Targeting this to Responsible Disclosure if it hasn't been caught already.

Found 13 items total: 1 critical, 0 high, 3 medium, 9 low.

---

exposed-secrets — relayer/.env.compose

Severity: CRITICAL

Language: Config

Environment file with potential secrets

---

Some of these might be false positives — just wanted to put them on your radar. Happy to provide more context if any of these look actionable.