Agentic SDLC is research into extending the Fisher loop across the software lifecycle: candidate mitigations, verification on isolated replicas, regression measurement, and change approvals informed by evidence.
Research & development · Not commercially available
Fisher finds and records failures, produces remediation guidance, and can re-test approved fixes where scoped. This work explores extending that loop into CI/CD, regression testing, and change approvals.
The proposed approach generates a candidate mitigation with its rationale, scope, and expected trade-offs.
The original exploit would be replayed and a fresh adaptive re-attack run against the remediated replica, graded on the replay result rather than a self-report.
The design intent is that production changes remain human-approved and informed by evidence.
Agentic SDLC is part of mace AI’s Deep Model Trust research and development. It is not commercially available.
Explore the research → Assess your current agent with Fisher