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/github-actions-supply-chain.json

View on GitHub
{
  "id": "github-actions-supply-chain",
  "title": "GitHub Actions supply-chain risk — third-party Action ingestion, SHA-pinning, pwn-request, cache poisoning",
  "description": "A platform engineer adopts third-party Actions in CI workflows. Recent incidents - trivy-action March 2026 (75/76 tags compromised), tj-actions/changed-files (~23k repos affected), Sysdig Shai-Hulud worm (November 2025), Wiz prt-scan campaign, and AI-scaled fork-PR pwn-request campaigns - have all exploited adjacent primitives: workflows referencing Actions by mutable tag, orgs without an allow-list policy, privileged workflows checking out untrusted PR code, and release/deploy jobs that expose secrets or OIDC tokens to code from the wrong trust boundary. Hulumi's `OrgFoundation` (M2) ships the SHA-pin-required Actions allowlist policy at startup-hardened tier; M3's `HulumiGithubHardeningPack` enforces this declaratively at preview-time; the workflow-governance linter gives consumers a local CI gate for workflow YAML drift.",
  "requiredFrameworks": [
    "CCM",
    "NIST-SSDF-v1.1",
    "OpenSSF-Scorecard",
    "MITRE-ATTCK",
    "GitHub-Well-Architected",
    "CIS-GitHub-v1.2.0"
  ],
  "actors": [
    "Platform Engineer (trusted, configures the org-level allowlist)",
    "Workflow Author (semi-trusted, picks Actions from the allow-listed set)",
    "Compromised Maintainer of a third-party Action (untrusted, primary attacker)",
    "External PR Attacker (untrusted, opens AI-generated fork PRs at scale)",
    "AI Coding Agent (trusted, authoring workflows; could be tricked into citing mutable tags)"
  ],
  "assets": [
    "Workflow YAML files referencing third-party Actions",
    "Organization-level Actions allowlist (`ActionsOrganizationPermissions`)",
    "Repository secrets accessible to the workflow",
    "OIDC token minting permission (`id-token: write`)",
    "Build artifacts produced by the workflow",
    "Cache entries in `actions/cache`"
  ],
  "stride": [
    {
      "type": "T",
      "name": "Third-party Action repo compromise (trivy-action / tj-actions pattern)",
      "description": "The maintainer's account is compromised (phishing, credential stuffing, or a separate supply-chain reach); attacker pushes a malicious commit to the Action's source repo and re-tags `v1` to point at it. Workflows referencing the Action by mutable tag pick up the malicious version on the next run. Mitigation: SHA-pin every external Action; org-level `ActionsOrganizationPermissions` policy with `shaPinningRequired: true` at startup-hardened tier.",
      "controls": [
        "CCM:DSI-04",
        "NIST-SSDF-v1.1:PO.3",
        "OpenSSF-Scorecard:Pinned-Dependencies",
        "MITRE-ATTCK:T1195",
        "GitHub-Well-Architected:PW.5"
      ]
    },
    {
      "type": "S",
      "name": "Cache poisoning via Action with cache-write capability",
      "description": "A compromised Action with `actions/cache` write access poisons the cache for a subsequent legitimate run. The legitimate run consumes the poisoned cache and ships compromised artifacts. Mitigation: cache scoping by branch + workflow; scope-limit Actions with cache-write to the minimum branch set.",
      "controls": [
        "CCM:DSI-02",
        "NIST-SSDF-v1.1:PW.6",
        "MITRE-ATTCK:T1195",
        "GitHub-Well-Architected:PW.6"
      ]
    },
    {
      "type": "E",
      "name": "AI-scaled pwn-request via PR diff content reaching write context",
      "description": "An external contributor opens many convincing fork PRs and relies on a privileged `pull_request_target` or downstream `workflow_run` chain to check out attacker-controlled head code. Package lifecycle scripts, build steps, or tests then run with base-repo secrets, write-capable `GITHUB_TOKEN`, or OIDC minting permission and exfiltrate credentials or publish poisoned artifacts. Mitigation: keep `pull_request_target` metadata-only, never check out PR head code in a privileged workflow, keep PR CI on `pull_request` with `permissions: contents: read`, and reject cloud OIDC trust subjects for pull requests.",
      "controls": [
        "CCM:IAM-07",
        "NIST-SSDF-v1.1:PW.7",
        "OpenSSF-Scorecard:Token-Permissions",
        "GitHub-Well-Architected:PS.2"
      ]
    },
    {
      "type": "I",
      "name": "Information disclosure via runner exfil after Action compromise",
      "description": "A compromised Action running on the org's runner reads environment, secrets, or filesystem state and exfils to an attacker-controlled URL. Step-Security Maintained Actions or octo-sts-style short-lived token exchange limit blast radius.",
      "controls": ["CCM:DSI-03", "NIST-SSDF-v1.1:PO.5", "MITRE-ATTCK:T1195"]
    }
  ],
  "recommendedComponents": [
    {
      "name": "@hulumi/baseline.github.OrgFoundation",
      "availability": "Shipped in Hulumi v1.1.0 M2",
      "rationale": "Provisions the org-level `ActionsOrganizationPermissions` with `shaPinningRequired: true` at startup-hardened tier; configures the Actions allowlist (selectedActionsPatterns)."
    },
    {
      "name": "@hulumi/policies.github.HulumiGithubHardeningPack",
      "availability": "Shipped in Hulumi v1.1.0 M3",
      "rationale": "H1 declaratively rejects raw `github.Repository` not wrapped in `SecureRepository`; H2/H3/H4 enforce ruleset, OIDC, and tier-monotonicity invariants."
    },
    {
      "name": "@hulumi/drift.adapters.GithubWebhookFallbackAdapter",
      "availability": "Shipped in Hulumi v1.1.0 M4",
      "rationale": "Detects org-level allowlist drift via the `repository_ruleset` and `organization` webhook events; surfaces `tier-degraded` honestly for non-GHEC tiers."
    },
    {
      "name": "scripts/workflow-governance-lint.mjs",
      "availability": "Shipped in Hulumi v1.4.0; WF_PR_1 added after the 2026-05-25 pwn-request review",
      "rationale": "Fails privileged `pull_request_target` / `workflow_run` workflows that check out attacker-controlled head code, alongside SHA pinning, token-permission, CODEOWNERS, and protected-environment checks."
    }
  ]
}
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: 194kUpdated Claude CodeCodexVerifiedInstall