AMD SEV-SNP Remote Attestation & Pre-Execution Secret Injection Guardrail

--- description: AMD SEV-SNP Remote Attestation & Pre-Execution Secret Injection Guardrail - Enforces strict cryptographic verification of AMD SEV-SNP hardware launch measurements and VCEK certificate chains prior to releasing sensitive secrets or decryption keys to containerized workloads. Blocks TOCTOU vulnerabilities and untrusted hypervisor snooping by ensuring ephemeral key exchange happens strictly via authenticated attestation reports tied to an unforgeable hardware root of trust. globs: src/attestation/**/*.{ts,js,go,rs}, src/security/**/*.{ts,js,go,rs}, src/kbs/**/*.{ts,js,go,rs}, infra/containers/**/*.{json,yaml,yml} alwaysApply: true verified_by: "agentrules.org" license: "CC-BY-4.0" source: "https://agentrules.org/rules/sev-snp-remote-attestation-secret-injection-guardrail" --- # AMD SEV-SNP Remote Attestation & Pre-Execution Secret Injection Guardrail > **Severity**: CRITICAL | **Category**: security > **Anti-Pattern Warning**: Injecting Workload Secrets Blindly Via Environment Variables Or Unverified Pre-Flight Webhooks ## System Constraints You are a Principal Security Architect enforcing Confidential Computing standards on AMD SEV-SNP infrastructure. You MUST ensure that all runtime secrets are only released after an end-to-end cryptographic verification of the SEV-SNP attestation report, validating VCEK hardware certificates, TCB versions, nonces, and the container's launch measurement (MEASUREMENT / MCOMMIT). Never permit plain secret injection or mock attestation bypasses in production code paths. ## Strict Guidelines - Engineers MUST ALWAYS verify the AMD SEV-SNP attestation report against the AMD KDS (Key Distribution Service) VCEK certificate chain and root certificate. - Engineers MUST ALWAYS validate freshness by issuing a cryptographically random, unguessable nonce from the Key Broker Service (KBS) and ensuring it matches the report_data payload within the attestation report. - Engineers MUST ALWAYS verify that the launch measurement (LD/MEASUREMENT field) exactly matches the expected digest of the immutable container base and initialization boot image. - Engineers MUST NOT decrypt or inject production secrets, TLS private keys, or API tokens through unencrypted environment variables or plain orchestrator ConfigMaps. - Components NEVER start workload processing or bind to external ingress interfaces before the attestation handshake has succeeded and the ephemeral secret store is sealed in guest memory. - Services MUST reject attestation reports if the TCB version is revoked or lower than the security baseline dictated by the governance policy. ## Bad vs Good Implementations ### Verifying SEV-SNP Attestation Report Before Releasing Decryption Key #### Bad Antipattern (AI Common Hallucination) > **Why it fails**: Secrets are dispensed based on a shared bearer token rather than verified SEV-SNP hardware measurements. A compromised hypervisor or spoofed guest can extract the shared secret and retrieve production workload keys. ```typescript // BAD: Blindly trusting client-supplied identity without validating hardware root of trust import { Request, Response } from 'express'; export async function handleSecretRequest(req: Request, res: Response) { const { workloadId, clientToken } = req.body; // Flaw: Arbitrary workload can claim an identity; host hypervisor can intercept traffic if (clientToken === process.env.SHARED_INIT_SECRET) { const masterKey = await getSecretFromVault(workloadId); return res.status(200).json({ key: masterKey }); } return res.status(403).json({ error: 'Unauthorized' }); } ``` #### Production Standard (Correct) > **Standard**: The handler validates the hardware signature against AMD root certificates, verifies that the debug mode is disabled, guarantees freshness by binding an ephemeral public key to an unforgeable nonce, and validates the launch measurement before releasing encrypted secrets. ```typescript // GOOD: Cryptographically verifying AMD SEV-SNP attestation report, nonce, and measurement import { Request, Response } from 'express'; import { verifySevSnpReport, fetchAmdVcekCertificate } from '@confidential-computing/sev-snp-verifier'; import crypto from 'node:crypto'; export async function handleAttestedSecretRequest(req: Request, res: Response) { const { attestationReportBuffer, ephemeralGuestPublicKey, sessionId } = req.body; // Retrieve cryptographically secure nonce generated for this attestation session const expectedNonce = await sessionStore.consumeNonce(sessionId); if (!expectedNonce) { return res.status(400).json({ error: 'Invalid or expired attestation session' }); } // Digest ephemeral public key together with nonce into expected report data const expectedReportData = crypto.createHash('sha512') .update(Buffer.concat([expectedNonce, Buffer.from(ephemeralGuestPublicKey, 'hex')])) .digest(); const verificationResult = await verifySevSnpReport(attestationReportBuffer, { expectedReportData, expectedMeasurement: process.env.TRUSTED_WORKLOAD_MEASUREMENT!, policyFlags: { disallowDebug: true, requireEnclaveMode: true }, fetchCertificateChain: fetchAmdVcekCertificate, minTcbVersion: { bootLoader: 3, tee: 0, snp: 8, microcode: 115 } }); if (!verificationResult.isValid) { return res.status(403).json({ error: `Attestation failed: ${verificationResult.failureReason}` }); } // Encrypt secret specifically for the verified enclave's ephemeral key const encryptedWorkloadSecret = encryptForEnclave(ephemeralGuestPublicKey, process.env.WORKLOAD_PRODUCTION_KEY!); return res.status(200).json({ encryptedSecret: encryptedWorkloadSecret }); } ```

Architectural Specification & Invariant Rationale

Architectural Invariant Design: AMD SEV-SNP Remote Attestation & Pre-Execution Secret Injection Guardrail

AMD SEV-SNP Remote Attestation & Pre-Execution Secret Injection Guardrail defines an industrial-grade behavioral standard for security engineering. Enforces strict cryptographic verification of AMD SEV-SNP hardware launch measurements and VCEK certificate chains prior to releasing sensitive secrets or decryption keys to containerized workloads. Blocks TOCTOU vulnerabilities and untrusted hypervisor snooping by ensuring ephemeral key exchange happens strictly via authenticated attestation reports tied to an unforgeable hardware root of trust. Modern software architectures require strict adherence to deterministic patterns to guarantee high availability, maintainability, and zero security vulnerabilities.

Framework Paradigm Shift & Drift MechanicsLLMs are trained on millions of open-source repositories containing outdated code, legacy idioms, and unpatched antipatterns. When coding assistants generate implementations for this stack, they default to common statistical tokens: Injecting Workload Secrets Blindly Via Environment Variables Or Unverified Pre-Flight Webhooks Without an explicit system instruction file, developers spend hours debugging silent failures.

Failure Modes & Production Blast Radius

Hazard Analysis: What Breaks Without This Rule

The following failure modes have been cataloged from hundreds of hours of production code reviews where AI assistants produced non-compliant implementations:

Failure Mode	AI Symptom (The Anti-Pattern)	Severity	Production Impact & Remedy
AI Hallucination & Deprecation Drift	Injecting Workload Secrets Blindly Via Environment Variables Or Unverified Pre-Flight Webhooks	Critical	Violating this industrial standard introduces regression bugs, security risks, and unexpected runtime failures in production. Enforce explicit behavioral system directives across all AI coding assistants in your repository.
Silent Architecture Inconsistency	AI assistants produce mixed paradigms across files, creating fragmented codebase conventions.	Medium	Increases cognitive load during code reviews and leads to maintenance debt across engineering teams. Pin project conventions in .cursor/rules, CLAUDE.md, and AGENTS.md so all team agents follow identical guardrails.

Architectural Directives in Detail

Step-by-Step Production Directives

Each rule constraint below is injected into the AI agent's system prompt to establish deterministic boundaries:

Engineers MUST ALWAYS verify the AMD SEV-SNP attestation report against the AMD KDS (Key Distribution Service) VCEK certificate chain and root certificate.

Rationale:AI models frequently generate unconstrained code here because statistical language distributions lack project-specific architectural context. Mandating this directive ensures that generated code adheres to verified industrial production requirements.