MAL-2026-5142
Vulnerability from ossf_malicious_packages
Published
2026-06-01 00:00
Modified
2026-06-01 00:00
Summary
Malicious code in @redhat-cloud-services/insights-client (npm)
Details

Part of the "Mini Shai-Hulud" supply chain worm campaign that compromised the GitHub Actions OIDC trusted publisher shared by Red Hat Cloud Services npm packages. The attacker injected a preinstall hook into this and 31 other packages in the @redhat-cloud-services scope. The hook delivers a three-layer obfuscated payload (ROT-9 Caesar cipher over a 1.27M-entry character-code array -> AES-128-GCM decryption with hardcoded keys -> stacked obfuscator.io encoding with PBKDF2+SHA-256 keystream S-box substitution) that downloads a pinned Bun runtime (v1.3.13) from GitHub to execute the worm outside the victim's Node installation.

Credential theft: Harvests AWS credentials (IMDS, ECS, Secrets Manager, SSM), Azure managed identities, GCP service account tokens, HashiCorp Vault tokens, Kubernetes service account tokens (/var/run/secrets/kubernetes.io/serviceaccount/token), GitHub PATs, npm publish tokens, environment variables from ~40 CI platforms (CircleCI, Travis CI, Jenkins, and others), password manager stores (Bitwarden, gopass), and local files (~/.npmrc, ~/.netrc, shell history, database history). Collected data is exfiltrated to attacker-controlled public GitHub repositories.

Privilege escalation: Exploits Docker socket access to escape containers and modify /etc/sudoers.d, granting passwordless sudo to CI runner user accounts.

Self-propagation: Uses stolen npm credentials to republish tampered tarballs of target packages. Injects a malicious CodeQL workflow into accessible GitHub repositories via the GraphQL createCommitOnBranch mutation, exchanges GitHub Actions OIDC tokens for npm publish tokens, and signs the resulting artifacts through Sigstore (Fulcio/Rekor) to appear legitimate.

Persistence and evasion: Installs a daemon at /tmp/kitty-<random>, hijacks .claude/settings.json for AI agent persistence, and hijacks .vscode/tasks.json for editor task execution. Detects sandbox environments via __FAKE_PLATFORM__, TESTING_TAR_FAKE_PLATFORM__, and __IS_DAEMON environment variables, and probes for EDR tools (CrowdStrike, SentinelOne, Carbon Black, StepSecurity Harden-Runner).

Credits
SafeDep safedep.io

{
  "affected": [
    {
      "package": {
        "ecosystem": "npm",
        "name": "@redhat-cloud-services/insights-client"
      },
      "versions": [
        "4.0.4",
        "4.0.5",
        "4.0.7"
      ]
    }
  ],
  "credits": [
    {
      "contact": [
        "https://safedep.io"
      ],
      "name": "SafeDep",
      "type": "FINDER"
    }
  ],
  "database_specific": {
    "malicious-packages-origins": null
  },
  "details": "Part of the \"Mini Shai-Hulud\" supply chain worm campaign that compromised the GitHub Actions OIDC trusted publisher shared by Red Hat Cloud Services npm packages. The attacker injected a `preinstall` hook into this and 31 other packages in the `@redhat-cloud-services` scope. The hook delivers a three-layer obfuscated payload (ROT-9 Caesar cipher over a 1.27M-entry character-code array -\u003e AES-128-GCM decryption with hardcoded keys -\u003e stacked obfuscator.io encoding with PBKDF2+SHA-256 keystream S-box substitution) that downloads a pinned Bun runtime (v1.3.13) from GitHub to execute the worm outside the victim\u0027s Node installation.\n\n**Credential theft:** Harvests AWS credentials (IMDS, ECS, Secrets Manager, SSM), Azure managed identities, GCP service account tokens, HashiCorp Vault tokens, Kubernetes service account tokens (`/var/run/secrets/kubernetes.io/serviceaccount/token`), GitHub PATs, npm publish tokens, environment variables from ~40 CI platforms (CircleCI, Travis CI, Jenkins, and others), password manager stores (Bitwarden, gopass), and local files (`~/.npmrc`, `~/.netrc`, shell history, database history). Collected data is exfiltrated to attacker-controlled public GitHub repositories.\n\n**Privilege escalation:** Exploits Docker socket access to escape containers and modify `/etc/sudoers.d`, granting passwordless sudo to CI runner user accounts.\n\n**Self-propagation:** Uses stolen npm credentials to republish tampered tarballs of target packages. Injects a malicious CodeQL workflow into accessible GitHub repositories via the GraphQL `createCommitOnBranch` mutation, exchanges GitHub Actions OIDC tokens for npm publish tokens, and signs the resulting artifacts through Sigstore (Fulcio/Rekor) to appear legitimate.\n\n**Persistence and evasion:** Installs a daemon at `/tmp/kitty-\u003crandom\u003e`, hijacks `.claude/settings.json` for AI agent persistence, and hijacks `.vscode/tasks.json` for editor task execution. Detects sandbox environments via `__FAKE_PLATFORM__`, `TESTING_TAR_FAKE_PLATFORM__`, and `__IS_DAEMON` environment variables, and probes for EDR tools (CrowdStrike, SentinelOne, Carbon Black, StepSecurity Harden-Runner).",
  "id": "MAL-2026-5142",
  "modified": "2026-06-01T00:00:00Z",
  "published": "2026-06-01T00:00:00Z",
  "references": [
    {
      "type": "REPORT",
      "url": "https://safedep.io/redhat-cloud-services-hit-by-mini-shai-hulud-npm-worm/"
    }
  ],
  "schema_version": "1.7.4",
  "summary": "Malicious code in @redhat-cloud-services/insights-client (npm)"
}


Log in or create an account to share your comment.




Tags
Taxonomy of the tags.


Loading…

Loading…

Loading…

Forecast uses a logistic model when the trend is rising, or an exponential decay model when the trend is falling. Fitted via linearized least squares.

Sightings

Author Source Type Date Other

Nomenclature

  • Seen: The vulnerability was mentioned, discussed, or observed by the user.
  • Confirmed: The vulnerability has been validated from an analyst's perspective.
  • Published Proof of Concept: A public proof of concept is available for this vulnerability.
  • Exploited: The vulnerability was observed as exploited by the user who reported the sighting.
  • Patched: The vulnerability was observed as successfully patched by the user who reported the sighting.
  • Not exploited: The vulnerability was not observed as exploited by the user who reported the sighting.
  • Not confirmed: The user expressed doubt about the validity of the vulnerability.
  • Not patched: The vulnerability was not observed as successfully patched by the user who reported the sighting.

Loading…

Detection rules are retrieved from Rulezet.

Loading…

Loading…