IronClaw — agentic threat model
IronClaw exhibits a high-risk agentic profile due to its ability to execute background routines, build dynamic WASM tools, and access IT/cybersecurity environments, but this is heavily counterbalanced by its robust, security-first architecture featuring Rust, Docker/WASM sandboxing, and local-first data privacy.
OWASP AIVSS score rationale
| Autonomy of Action | 0.80 | |
| Goal-Driven Planning | 0.70 | |
| Self-Modification | 0.60 | |
| Dynamic Tool Use | 0.80 | |
| Persistent Memory | 0.70 | |
| Contextual Awareness | 0.60 | |
| Dynamic Identity | 0.40 | |
| Multi-Agent Interactions | 0.30 | |
| Non-Determinism | 0.50 | |
| Opacity & Reflexivity | 0.50 |
Scored with the canonical OWASP AIVSS formula (AIVSS calculator reference); agentic risk factors estimated from the agent’s described capabilities.
MAESTRO 7-layer threat model
Per-layer threats for this agent. Layers tagged “not certain from listing” are general, caveated commentary where the public description didn’t pin that layer.
Not certain from the listing — The description does not specify which foundation models are bundled or supported locally vs. via API, though it mentions protecting against prompt injection.
IronClaw prioritizes local storage, encrypted secrets, and zero telemetry, significantly reducing the risk of external data exfiltration and unauthorized knowledge-base exposure.
Supports background routines, parallel jobs, and dynamic WASM tool building. While powerful, these orchestration features are mitigated by local control and sandboxing.
Strong infrastructure security posture utilizing Rust, Docker sandboxing, and WASM sandboxing to isolate dynamic tools, alongside multi-channel gateways (Slack, Telegram, HTTP webhooks).
Not certain from the listing — Mentions heartbeat-based monitoring for background routines, but does not detail comprehensive evaluation, guardrails, or observability logging frameworks.
Designed with a security-first approach, featuring encrypted secrets, local-only data retention, and layered defenses specifically targeting prompt injection and data exfiltration.
Not certain from the listing — While it supports MCP (Model Context Protocol) servers and dynamic tools, it does not explicitly detail multi-agent orchestration or marketplace interactions.
MAESTRO — the 7-layer agentic threat-modeling framework (Cloud Security Alliance / Ken Huang).
These scores are auto-generated from public information (the agent's own listing, docs, and repository) using the canonical OWASP AIVSS formula and the MAESTRO framework — an estimate for guidance, not a penetration test, audit, or certification. See the scoring methodology. Are you the vendor? Factual corrections are free.