AI penetration-testing vendor pairs its own liability insurance with a money-back delivery guarantee

Agent shield, an AI-driven penetration-testing service marketed at agentshield.win under the name Tough Love Security, has been added to AI Dispatch's index. The vendor runs a pipeline it describes as reconnaissance, attack-surface mapping, vulnerability analysis, conditional exploitation and reporting, aimed at web applications and APIs. Its own FAQ states: "Every finding is reviewed by a senior security engineer before delivery. The AI agents do the heavy lifting — recon, data flow analysis, payload generation, exploitation — but a human validates each finding, removes false positives, and writes remediation guidance."

On liability, the company states: "We carry professional liability insurance." It names no carrier and no coverage limit, and separately states: "Our liability is limited to the engagement fee paid." Pricing runs from a free scan to an Enterprise tier at "$10K+ annual contract," with a flat-fee HIPAA package — "$3,500 Flat" — aimed at medical, dental and behavioral-health practices. Delivery is backed by a guarantee: "10-day delivery. 30-day money-back if we miss."

Assessments are contractually bounded. Terms of service state: "Testing is conducted exclusively within the boundaries defined in the testing agreement. Any systems, endpoints, or networks not explicitly listed in the scope document will not be tested." The company says it will not begin without authorization: "All assessments require explicit written authorization from the target owner. We do not test systems without proper authorization under any circumstances."

AI Dispatch scores agents and vendors on a five-dimension trust-gap rubric: runtime governance, audit trail, compliance, outcome settlement, and insurance/indemnity. Agent shield scores 1 of 3 on both insurance/indemnity and outcome settlement — the only entity, among the nine AI insurance and brokerage vendors AI Dispatch has indexed plus ten other fully audited agent and governance vendors, to record evidence on both dimensions at once. All nine insurers score 0 on outcome settlement, and none of the governance and agent vendors previously scored — Holistic AI, Credo AI, Langfuse and gptme among them — scores above 0 on insurance/indemnity. Agent shield's own coverage remains unnamed and uncapped in public dollar terms, and its guarantee ties a refund to a missed delivery date, not to a verified security outcome.

Full field-level sourcing for Agent shield is published at aidispatch.news.