-
Notifications
You must be signed in to change notification settings - Fork 0
AI Red Team And Supply Chain Gates
Huzefaaa2 edited this page Jul 4, 2026
·
1 revision
CAVRA R6.3 adds native AI red-team and AI supply-chain gates. The public contract validates LLM guardrail tests, AI artifact supply-chain metadata, malicious model indicators, and red-team closeout evidence without moving raw prompts, model weights, training data, private features, or customer records.
sequenceDiagram
participant Test as Guardrail tests
participant Scan as Supply-chain scan
participant Model as Malicious model checks
participant Gate as CAVRA AI red-team gate
Test->>Gate: prompt injection and unsafe request results
Scan->>Gate: digest, provenance, SBOM, dependencies
Model->>Gate: serialization, remote code, hidden payload checks
Gate-->>Gate: require no blockers
prompt_injection_overridesecret_exfiltration_requestunsafe_tool_chain_requestdata_export_without_scope
- artifact digest
- provenance reference
- SBOM reference
- serialization safety
- dependency allowlist
- no raw model egress
- unsafe serialization
- remote code execution
- hidden prompt payload
- dependency confusion
cavra ai-red-team guardrails
cavra ai-red-team supply-chain --artifact examples/ai-red-team/ai-artifact-metadata.sample.json
cavra ai-red-team malicious-model --artifact examples/ai-red-team/ai-artifact-metadata.sample.json
cavra ai-red-team export --output-dir dist/ai-red-team
cavra ai-red-team readiness examples/ai-red-team/enterprise-ai-red-team.live.sanitized.example.json --require-livesrc/cavra/ai_red_team.pyscripts/validate_ai_red_team.pyexamples/ai-red-team/guardrail-test-suite.sample.jsonexamples/ai-red-team/ai-artifact-metadata.sample.jsonexamples/ai-red-team/ai-artifact-metadata.invalid.jsonexamples/ai-red-team/enterprise-ai-red-team.sample.jsonexamples/ai-red-team/enterprise-ai-red-team.live.sanitized.example.json.github/workflows/ai-red-team.ymltests/test_ai_red_team.py
The live gate is accepted only when:
{
"ready_for_live_ai_red_team_gate": true,
"blocker_count": 0
}
CAVRA Field Compass
Before the agent acts, CAVRA asks: who is acting, what will change, what policy applies, and what evidence will prove it?
Before the agent acts, CAVRA asks: who is acting, what will change, what policy applies, and what evidence will prove it?
| Start | Build | Operate | Assure |
|---|---|---|---|
| Quick Start | CLI | Enterprise Guide | AISPM |
| Reader Paths | Policy Syntax | Deployments | Evidence |
| Community | GUI | Troubleshooting | Conclusion |
Textbook home: Before the Agent Acts |
Development archive: development and testing artifacts |
Source repository: github.com/Huzefaaa2/cavra
- Foreword And Reader Paths
- Why CAVRA Exists
- Runtime Authority Model
- Architecture
- Editions
- Install And Deploy
- Community Guide
- Enterprise Guide
- CLI Reference
- GUI And Sandbox
- AISPM Guide
- Policy And Evidence
- Operations And Integrations
- Labs And Use Cases
- Appendices And FAQ
- Policy Language Reference
- Troubleshooting Playbook
- Technology Stack
- Unified Enterprise Roadmap
- Conclusion
- Development And Testing Archive
- Unified Enterprise Roadmap
- CLI
- API
- CAVRA Trial Field Guide
- AISPM Enterprise Live Ingestion
- Enterprise HA/DR Readiness
- Enterprise HA/DR Azure Map
- Enterprise KMS/HSM Evidence Custody
- Enterprise Immutable Audit Log
- Enterprise Compliance Mapping Packs
- Enterprise Reporting Exports
- Connector SDK And Certification
- Priority Certified Connectors
- Model Registry Connectors