CVE-2023-33733: Reportlab vulnerable to remote code execution

The vulnerability stems from Reportlab's HTML-to-PDF processing using a sandboxed eval environment (rl_safe_eval). Key indicators:

1. restricted_eval is the entry point for evaluating untrusted expressions in markup attributes
2. rl_getattr's security checks were bypassed using a crafted string subclass
3. rl_type restrictions were circumvented to access native type constructor These functions would appear in stack traces when processing malicious color attributes containing exploit payloads. The exploit writeup demonstrates direct manipulation of these sandbox components to achieve RCE.