Skip to content

kerberosmansour/hulumi / skills/hulumi-threat-model

hulumi-threat-model

Flagged

Run an AWS or GitHub cloud / platform threat model for a specific Hulumi scenario. Produces a structured markdown threat-model document at docs/threat-model-<scenario>-<date>.md in the user's working directory, citing CSA CCM v4.1, NIST SP 800-53 Rev 5, NIST SP 800-218A, MITRE ATLAS v5.1, CIS AWS Foundations v5.0.0, CIS GitHub Benchmark v1.2.0, NIST SSDF v1.1, OpenSSF Scorecard, MITRE ATT&CK T1195, and GitHub Well-Architected SSDF by framework ID only (no verbatim control text). Use when the user asks for a threat model for an AWS or GitHub scenario, wants to ground IaC design decisions in standard controls, or wants to see which @hulumi/baseline components apply.

scenarios/operations-patch-compliance-lapse.json

View on GitHub
{
  "id": "operations-patch-compliance-lapse",
  "title": "Operations — Patch Compliance Lapse via Silent Un-Patching",
  "description": "An engineer (or AI agent) configures `Ec2PatchBaseline` with `rebootOption: { kind: \"NoReboot\" }` at startup-hardened tier, intending to avoid disrupting tenants. The component refuses construction (M7 invariant), but the threat path remains real: a sandbox-tier consumer might accept silent un-patching, leaving a kernel CVE exposed for sixty days. Target end state matches Hulumi's `Ec2PatchBaseline` (M7) + `Ec2PatchWaves` (M7) + `HulumiOperationsHardeningPack`'s `O_PATCH_1` (M10). Threats focus on the gap between an EC2 patch announcement and the cluster reboot that activates it.",
  "requiredFrameworks": ["CCM", "NIST-800-53-r5", "ATLAS", "CIS-AWS-v5.0.0"],
  "actors": [
    "Platform Engineer (trusted)",
    "AI Coding Agent (trusted, but may suggest NoReboot for tenant-stability reasons)",
    "External Attacker (post-patch-announcement, racing to exploit before reboot)"
  ],
  "assets": [
    "EC2 instance kernel and userspace package state",
    "SSM Patch Compliance metric (per Patch:Group)",
    "Patch:Group tag (load-bearing for wave membership)"
  ],
  "stride": [
    {
      "type": "T",
      "name": "Tampered Patch:Group tag bypasses the wave-composer",
      "description": "An attacker (or a misconfigured IaC author) sets a Patch:Group value outside {dev, staging, production} on an EC2 instance. The MaintenanceWindowTarget no longer matches; the instance is silently excluded from patching. Mitigation: `O_PATCH_1` (M10) rejects the tag value at preview; SCP-style guardrails on `ec2:CreateTags` for that tag complement at runtime.",
      "controls": ["CCM:CEK-08", "CIS-AWS-v5.0.0:1.20", "NIST-800-53-r5:CM-2"]
    },
    {
      "type": "R",
      "name": "Repudiation via missing compliance metric",
      "description": "Without `complianceMetric.topicArn` wired to a real SNS, patch-compliance regressions surface as a CloudWatch metric nobody reads. Mitigation: `Ec2PatchBaseline` requires the SNS ARN (M7 invariant); the alarm fires when non-compliant instance count >= 1.",
      "controls": ["CCM:LOG-01", "NIST-800-53-r5:AU-6"]
    },
    {
      "type": "I",
      "name": "Information disclosure via NoReboot pile-up",
      "description": "A `NoReboot` posture at startup-hardened tier means kernel CVEs accumulate without activation. `Ec2PatchBaseline` refuses this combination at construction (M7 type system); the sandbox-only escape requires a written `hulumi_decision_comment` (>= 8 chars) tagged on the resource for audit.",
      "controls": ["CCM:VUL-01", "NIST-800-53-r5:RA-5", "ATLAS:AML.T0001"]
    },
    {
      "type": "D",
      "name": "Denial via synchronized-reboot storm",
      "description": "Without CRC32-bucket staggering, all instances reboot at the maintenance window's start, taking the service offline simultaneously. Mitigation: `staggering.bucketCount` defaults to 4 and is bound at 16; max-concurrency on the run-command task scales inversely with bucket count.",
      "controls": ["CCM:UEM-09", "NIST-800-53-r5:CP-2"]
    },
    {
      "type": "E",
      "name": "Elevation via unpatched root-exploit kernel",
      "description": "An unpatched local-privilege-escalation kernel exploit lets a compromised non-root process become root. Sixty days is the empirical compromise window. Mitigation: tier-aware patch approval delay (3 days at startup-hardened, 7 at sandbox).",
      "controls": ["CCM:VUL-01", "ATLAS:AML.T0001"]
    }
  ],
  "abuseCases": [
    {
      "id": "tm-hulumi-ops-abuse-noreboot-without-decision",
      "title": "NoReboot at startup-hardened tier",
      "description": "Mitigation: M7 constructor refuses the combination.",
      "controls": ["CCM:VUL-01"]
    },
    {
      "id": "tm-hulumi-ops-abuse-patch-group-tag-tamper",
      "title": "Free-form Patch:Group tag bypasses wave-composer",
      "description": "Mitigation: `O_PATCH_1` (M10) at preview.",
      "controls": ["CCM:CEK-08"]
    }
  ]
}
rea-analysis-optimizationmorluto/reaOptimize REA providers and shared analysis workflows for agent context, latency, CPU, and memory while preserving correctness and Evidence. Use for analysis bottlenecks and cache or query design; CI/test-suite optimization belongs to rea-ci-tests-audit.Stars: 472kUpdated Claude CodeCodexVerifiedInstall
rea-ci-tests-auditmorluto/reaAudit REA CI and test suites for correctness, fixture drift, false passes, and measured performance; use for suite or workflow investigations and efficiency changes, rather than ordinary feature testing or a single test failure.Stars: 472kUpdated Claude CodeCodexVerifiedInstall
rea-tool-designmorluto/reaAudit, investigate, design, or change REA investigation tools, CLI/MCP contracts, provider capabilities, and Evidence semantics.Stars: 472kUpdated Claude CodeCodexVerifiedInstall
ponytail-auditDietrichGebert/ponytailQuality audit of the whole repo: bugs, security, real load, missing tests, speed, and what to delete. Most important first.Stars: 196kUpdated Claude CodeCodexVerifiedInstall