Case studies
These case studies show how persistent MindStone Agents use layered continuity, durable memory, verification discipline, and domain expertise to do serious long-running work.
MindStone agent team · Claude Code / agent orchestration
Four live MindStone-Agent experiments comparing Claude Fable 5 and Claude Opus 4.8 across codebase-grounded design and a greenfield app: Fable led on details, and MindStone/LCA review workflow moved Opus into the Fable 5 quality band.
MindStone Agent incident-response agent · Ransomware response / cyber recovery
A persistent MindStone Agent helped turn a full-server ransomware compromise into a zero-data-loss, zero-ransom incident response: forensic preservation, clean recovery, hardening, and production cushion cutover in a single working day.
Hearth · DevOps / Infrastructure migration
Hearth migrated five interdependent production services off a post-ransomware recovery server in about two hours, with zero downtime, zero data loss, CI/CD modernization, infrastructure hardening, and legacy cloud spend eliminated.
Aegis · OT/ICS cyber threat intelligence
Aegis operates as a persistent OT/ICS threat analyst: running daily intelligence pulls, maintaining threat actor lifecycle state, enriching advisories, managing KEV deadlines, and publishing dashboard-ready intelligence output.
Format
Each case study has a public web page and downloadable PDF collateral when available.
What problem or workflow the case study covers, and why continuity or review discipline mattered.
Which layers were involved: identity, transcript continuity, structured memory, recall, Synapse, or checkpointing.
What changed: speed, accuracy, handoff quality, drift reduction, collaboration, or artifact quality.
Download links point to generated PDF versions stored in the public assets tree.