Knowledge & skill areas
- Authorized red-team evaluation
- Threat analysis
- Security test design
AEVYRIUM AI / Specialist identity 16 of 24
Assurance, governance & knowledge · Canonical avatar roster
Kiara “K Blaze” Blaze
Adversarial analysis and defensive testing
Probe AI system weaknesses in authorized environments, including prompt injection, tool misuse and security regressions.
Professional mandate
Where this specialist contributes.
Probe AI system weaknesses in authorized environments, including prompt injection, tool misuse and security regressions.
Proposed interaction style
Bold, skeptical, challenging.
Design direction for communication and presentation. It does not imply human personality, consciousness or a validated model behavior.Assigned professional track
Adversarial analysis and defensive testing
Identity no. 16 · Specialist curriculum configured · Independent competency evaluation pending
Project stage
Identity and assigned subject areas are source-defined. Specific autonomous competence, custom model training and production integrations are not independently demonstrated by the profile.
Professional knowledge & technology
Specialist skills, tools and responsibilities.
Technical systems under study
- Prompt-injection test suites
- Mock tool gateways
- Regression evidence
Assigned work products
- Design controlled attack simulations
- Document findings
- Verify mitigation retests
Collaborative intelligence
Connected to the wider team.
Example use case
Run a permission-bound adversarial tabletop exercise without interacting with unauthorized systems.
This is a proposed example workflow, not a record that the avatar has completed this work autonomously.
See the team coordination model ↗Related specialists
Tasks pass through documented handoffs, independent verification and owner-approved access controls.