Guardrail Regression

Two guardrail configurations are on the table and somebody has to choose. Run twenty trials against each with the same harness, then make a recommendation you can defend on both blocked attacks and broken legitimate work.

Advanced · Several moving parts — expect trade-offs and a wrong turn or two.

Outcome · Raise attack resistance while retaining measured legitimate utility.

Launch the linux desktop Open the workbench only How this is graded

George · 5 lines

Roles

• AI security engineer

Standards mapping

NICE T0177

ATLAS AML.T0043, AML.T0054

OWASP LLM01, LLM09

Publish gate

Manifest valid · release 1.0.0

AI connection

This activity runs entirely on the deterministic runtime. Optional live-model runtimes only affect dialog, coaching and wording — never checks or scoring.

Deterministic simulationmode 0

Nothing leaves the browser session.

Chrome built-in AImode 1

Prompts never leave the learner's machine.

Bring your own keymode 2

Lab prompts reach the learner's chosen provider only.

Organization-owned AImode 3

Customer tenancy; Cybrary sees metadata only.

Cybrary-managed AImode 4

Cybrary AI gateway; retention off by default.

Hybrid routingmode 5

Per-job — the router never sends restricted jobs off-tenant.

Lab manifest

The machine-readable release contract. Authors edit this; the console refuses to publish without checks, both regression suites and a reference solve.

schemaVersion: cybrary.ai-lab/v1
id: guardrail-regression
release: 1.0.0
mode: ai-lab
maturity: 3
roles:
  - ai-security-engineer
outcomes:
  niceTasks:
    - T0177
  atlasTechniques:
    - AML.T0043
    - AML.T0054
  owaspRisks:
    - LLM01
    - LLM09
aiAssets:
  models:
    - support-agent-model@2026.06
  prompts:
    - guardrail-regression-system-prompt@1
  agents:
    - support-agent-orchestrator@3
  tools:
    - crm.lookup
    - ticket.update
    - email.send
    - refund.issue
  corpora:
    - support-kb@v4
enterpriseTopology: support-enterprise-v3
scenario:
  seed: guardrail-regression-seed-1
  injects:
    - guardrail-regression-inject-1
telemetry:
  requiredSignals:
    - ai.trace
    - rag.retrieval
    - tool.call
    - identity.auth
    - dlp.egress
    - siem.alert
evaluation:
  deterministicChecks:
    - guardrail-regression-checks
  repeatedTrials: 20
  maliciousSuite: guardrail-regression-attack-v1
  benignSuite: support-utility-v4
safety:
  networkEgress: deny
  maxTokens: 250000
  maxCostUsd: 20