Free OpenClaw security scanner. 3,000+ agents audited. 3-Layer Audit Protocol. OWASP ASI 10/10 coverage. AI agent integrity layer.

🔍 Auditoria de segurança defensiva para Agent Skills (Claude, OpenClaw). Detecta payload em archives, ASCII smuggling, supply chain attacks e meta-prompt injection. Alinhado a OWASP ASI 2026 + AST01. Camada 1 da defesa em profundidade.

Reference reproducer for the Tool Output Mimicry primitive (Brana 2026, doi:10.5281/zenodo.19794072) — bypasses multi-layer agentic AI defenses via upstream-agent impersonation in user-controllable fields.

Policy-gated AI decisions with NIST FIPS 204 post-quantum receipts. EU AI Act Article 50 native. Every MCP tool call signed, auditable, provable offline.

AgentGuard benchmark suite — 56 hand-crafted samples validating all 22 detection rules

AgentGuard VS Code extension — inline agent security diagnostics as you type

AgentGuard Demo — CI/CD integration with 8 real Code Scanning alerts on every push

AgentGuard GitHub App — automated PR security scanning with SARIF upload

Pre-agent metadata scanner. Catches prompt injection in branch names, commits, file names, PR titles, and other untrusted strings before they reach an AI code reviewer.