All 100 adversarial probes targeting prompt boundaries, network egress, shell escape, and tool authorization were neutralized in-line by Tcha Protection defense engines.
| Attack Domain | OWASP LLM Matrix | Probes | Neutralized | Status |
|---|---|---|---|---|
| Indirect Prompt Injection & Steganography | LLM01:2025 Prompt Injection | 25 | 25 | ✓ 100% PASS |
| SSRF & Network Egress Exfiltration | LLM02:2025 Sensitive Info / SSRF | 25 | 25 | ✓ 100% PASS |
| Path Traversal & Shell Metacharacters | LLM05:2025 Improper Output Handling | 25 | 25 | ✓ 100% PASS |
| Tool Abuse & Memory Poisoning | LLM06:2025 Excessive Agency & Autonomy | 25 | 25 | ✓ 100% PASS |
| Standard / Regulation | Requirement Reference | Tcha Control Enforced | Compliance Status |
|---|---|---|---|
| EU AI Act (Regulation 2024/1689) | Article 15: Cybersecurity & Robustness | Pre-execution prompt sanitization & automated red-team audit | COMPLIANT |
| SOC 2 Type II | CC6.1 / CC6.6 Boundary Protection | In-line SSRF egress filter, CIDR deny rules, IMDS blocking | COMPLIANT |
| OWASP Top 10 for LLMs 2025 | LLM01, LLM02, LLM05, LLM06, LLM07 | Multi-layer defense: Delimiter encapsulation, HITL gate, Merkle ledger | COMPLIANT |
| ISO/IEC 42001:2023 | A.6.2 AI System Risk Mitigation | Cryptographic Merkle hash-chained tamper-evident audit logging | COMPLIANT |