Skip to main content

Accelerated GenAI Security Assessment

A guided walkthrough of a complete Tessera ecosystem, start to finish. The Accelerated GenAI Security Assessment ecosystem reviews an agentic AI workload on AWS and produces a defensible security assessment — anchored on the AWS Well-Architected Agentic AI Lens (Security pillar, AGENTSEC01–09), cross-mapped to the OWASP agentic threat taxonomy (ASI01–10 + LLM01–10) and the CSA MAESTRO layered model, verified against MITRE ATLAS, and grounded in the durable cyber-security-risk catalog. It fails closed if that framework grounding is absent.

It runs in two phases:

  1. Assessment playbook — a design-level review of the target's architecture that produces the assessment kit: a scoped questionnaire, test procedures, a tooling guide, findings/report templates, an execution playbook, a local code-review SOP, and a design workshop deck. Findings here are hypotheses to confirm in phase 2.
  2. Findings reconciliation & report — consumes the evidence gathered by executing phase 1 (filled questionnaire, findings register, SAST/SCA SARIF, CSPM ASFF), reconciles it, scores risk, audits coverage, and emits a self-contained HTML report and stakeholder read-out deck, plus editable Word and PowerPoint renditions — with a humanizer and a final quality gate before emission.

The four short clips below follow the full loop: build the ecosystem in the visual Builder, run it, watch the live run, and review the report. Each video is click-to-play.

1. Build the ecosystem​

Compose the ecosystem in the visual Builder — the two-phase structure, the agents that map the AWS Agentic AI Lens and OWASP/MAESTRO/ATLAS frameworks, and the kit + report artifacts — then save and deploy. No code; the Builder compiles the agent graph into a runnable workflow. For the full walkthrough, see Build an ecosystem.

2. Run it​

Open the saved ecosystem, attach the target's architecture documentation (and, for phase 2, the gathered evidence), and launch a run.

3. Watch the live run​

The run streams in real time: specialist agents map the AWS AI attack surface and agentic threats, cross-walk the AGENTSEC controls, and assemble the assessment kit — with the DAG map showing progress across both phases.

4. Review the report​

When the run finishes, review the final assessment — the self-contained HTML report and stakeholder read-out deck, with findings reconciled against the gathered evidence, risk scored, and framework coverage audited. Downloadable artifacts include the editable Word and PowerPoint renditions and the full kit/report bundles. For how to read a run and iterate on it, see Run and review.

Try it yourself​

  1. Deploy and seed the platform (Installation).
  2. Open the UI and build or pick this ecosystem.
  3. Attach the target architecture, launch — then watch the run and review the report.