Инструменты

Orchestrates subagents in-session to map the architecture, build a threat model and hunt for vulnerabilities. Every finding is independently verified before it goes in the report, which also comes as SARIF, and optional patch files are left for you to apply. Official Anthropic plugin.

Three layers of security checks on agent-written code: rule-based warnings on about 25 risky patterns during edits, a model review of the diff at turn end, and an agent tracing cross-file bugs like IDOR and SSRF at commit. Official Anthropic plugin; needs Python 3.8+ and model API access.

Compliance skill from Anthropic's official operations plugin: maps controls, owners and evidence for SOC 2, ISO 27001, GDPR, HIPAA and PCI DSS, schedules audits, runs gap analyses, and outputs audit-readiness checklists and evidence collection plans. Have legal or compliance experts review.

Anthropic's official AI governance plugin: tiers AI use cases against your internal register, runs impact assessments across applicable laws such as the EU AI Act, reviews vendor AI terms for training-data and liability risk, and periodically checks for drift between AI policy and practice.

Privacy compliance plugin: decides whether processing activities need a PIA or DPIA, generates privacy impact assessments, reviews DPAs from the controller or processor side, and drafts DSAR responses within statutory deadlines. Official from Anthropic.

Anthropic's reference implementation for finding and fixing vulnerabilities: Claude Code skills for threat modeling, static scanning, triage and patching, plus an autonomous recon-to-patch pipeline (C/C++ memory bugs by default; needs Docker and gVisor). No longer maintained.

Anthropic's official OS-level sandbox CLI: restricts file and network access for agents, MCP servers and any command without containers, using sandbox-exec on macOS and bubblewrap on Linux. Currently a research preview.