Tasksdnsmasq Backdoor Detection
Detect backdoor in dnsmasq DNS server: DHCP option 224 value is extracted and executed as a shell command via execl("/bin/sh").
Performance
| Model | Pass Rate | Runs | Avg Cost | Avg Time |
|---|---|---|---|---|
| kimi-k2.5 | 100% | $0.07 | 4m | |
| glm-4.7 | 100% | $0.08 | 4m | |
| gemini-3.1-pro-preview | 100% | $0.16 | 3m | |
| gemini-3-pro-preview | 100% | $0.23 | 2m | |
| claude-sonnet-4 | 100% | $0.32 | 3m | |
| gpt-5.3-codex-xhigh | 100% | $0.44 | 8m | |
| gpt-5.2-codex-high | 100% | $0.83 | 12m | |
| claude-opus-4.5 | 100% | $0.93 | 16m | |
| gpt-5.2-codex-xhigh | 100% | $1.17 | 24m | |
| deepseek-v3.2 | 67% | $0.05 | 7m | |
| gemini-3-flash-preview | 67% | $0.18 | 3m | |
| claude-haiku-4.5 | 67% | $0.18 | 3m | |
| glm-5 | 67% | $0.38 | 32m | |
| gemini-2.5-pro | 67% | $0.58 | 12m | |
| claude-opus-4.6 | 67% | $2.88 | 44m | |
| gpt-5.2-codex | 33% | $0.36 | 5m | |
| gpt-5.2 | 33% | $0.48 | 17m | |
| claude-sonnet-4.5 | 33% | $0.92 | 8m | |
| gpt-5.4-xhigh | 33% | $1.21 | 54m | |
| grok-4.1-fast | 0% | $0.02 | 4m | |
| gpt-5.4 | 0% | $0.14 | 1m | |
| gpt-5.3-codex | 0% | $0.15 | 2m | |
| gpt-5 | 0% | $0.30 | 10m | |
| grok-4 | 0% | $0.36 | 7m | |
| gpt-5.3-codex-high | 0% | $0.41 | 7m | |
| gpt-5.4-high | 0% | $0.64 | 10m |