Lightning PyPI Bun-Based Credential Stealer

Executive Summary

On April 30, 2026, two malicious releases of the legitimate PyPI package lightning were published as versions 2.6.2 and 2.6.3 Lightning advisory Snyk Socket. The package is the modern distribution name for the PyTorch Lightning deep learning framework, making the compromise materially different from a lookalike or typosquat.

The malicious wheels added a hidden _runtime directory that executed when Python code imported lightning, downloaded the Bun JavaScript runtime from GitHub, and used it to run an approximately 11 MB obfuscated JavaScript credential stealer named router_runtime.js Snyk Socket. The payload targeted GitHub tokens, npm tokens, cloud credentials, metadata services, local environment variables, and developer credential files Socket. The project advisory tracks the incident as GHSA-w37p-236h-pfx3 / CVE-2026-44484 and states that the malicious releases were quarantined Lightning advisory.

Key Facts

Threat Type: legitimate package compromise, import-time credential stealer

Ecosystem: pypi, python

Registry: PyPI

Affected Packages:

• lightning

Malicious Versions:

• 2.6.2
• 2.6.3

Known Good Versions:

• 2.6.1

Post Incident Versions Not Listed As Affected:

• 2.6.4
• 2.6.5

Execution Trigger: import lightning

Primary Impact: Developer, CI/CD, npm, GitHub, and cloud credential theft

Campaign Context: Mini Shai-Hulud-style Bun payload reuse; exact attribution remains vendor-disputed

Known Iocs:

• lightning/_runtime/start.py
• lightning/_runtime/router_runtime.js
• github.com/oven-sh/bun/releases/download/bun-v1.3.13
• api.github.com/user
• registry.npmjs.org/-/whoami

Confidence: high

Canonical Source: hxxps://snyk[.]io/blog/lightning-pypi-compromise-bun-based-credential-stealer/

Evidence Assessment

• confirmed: lightning versions 2.6.2 and 2.6.3 were malicious and are covered by the project’s GHSA-w37p-236h-pfx3 / CVE-2026-44484 advisory Lightning advisory.
• confirmed: The malicious execution chain runs automatically on module import and launches a background process with suppressed output Snyk Sonatype.
• confirmed: The payload uses Bun to execute a large obfuscated JavaScript credential stealer that targets cloud and developer credentials Snyk Sonatype.
• confirmed: The project says the attacker captured PyPI credentials; the security advisory says the exact root cause remained under investigation in its May 12 update Lightning incident review Lightning advisory.
• confirmed: PyPI no longer exposes 2.6.2 or 2.6.3; current registry metadata lists 2.6.4 and 2.6.5, while the incident advisory still names 2.6.1 as its explicit fallback PyPI Lightning advisory.
• unclear: Vendors differ on whether the operator is the original Shai-Hulud actor, a copycat, or a related cluster reusing the same payload family Snyk.

Impact Determination

Classification	Criteria	Required evidence	Required action	Closure condition
Confirmed compromise	lightning==2.6.2 or lightning==2.6.3 is present and Python import-time loader starts Bun and obfuscated JavaScript or the reported process, file, or network indicators is observed.	Artifact inventory plus runtime telemetry showing Python import-time loader starts Bun and obfuscated JavaScript or listed C2/process/file indicators.	Isolate affected hosts or runners, preserve artifacts, and rotate reachable credentials from a clean environment.	Affected artifacts are removed, exposed credentials are replaced, and downstream audit modules show no suspicious follow-on use.
Presumed exposed	lightning==2.6.2 or lightning==2.6.3 was installed, pulled, imported, built, or executed during the exposure window, but telemetry cannot prove exfiltration.	Lockfile, package cache, workflow, image pull, extension inventory, build log, or deployment record tied to the exposure window.	Rebuild from clean artifacts and rotate credentials available to the affected environment.	Credential owners confirm revocation of old material and clean artifacts are deployed.
Potentially exposed	The package, workflow, image, extension, or module appears in dependency or deployment records, but Python install, import, or interpreter-startup execution is not established.	Manifest, lockfile, build, deployment, or endpoint records plus a named telemetry gap.	Collect the missing execution and telemetry evidence before narrowing scope.	Every hit is dispositioned as confirmed compromise, presumed exposed, or not exposed.
Not exposed	No affected version, artifact, mutable reference, or indicator appears in source, lockfiles, build outputs, deployments, package caches, or runtime telemetry.	Repository search, dependency inventory, build/deployment export, package cache query, and runtime telemetry query results.	Preserve the negative search output and keep the prevention controls active.	Search evidence covers developer endpoints, CI runners, production deployments, and package or image caches.
Unknown	Required inventory, build, endpoint, network, or audit telemetry is unavailable.	A gap statement naming unavailable systems, owners, and time windows.	Keep the asset in scope and make conservative rotation or rebuild decisions for high-value environments.	The missing evidence is recovered or the risk owner accepts residual uncertainty.

Minimum Evidence To Collect

Minimum Evidence:

• Dependency, workflow, extension, image, or module inventory covering developer endpoints, CI runners, and production deployments.
• Positive or negative search results for lightning==2.6.2, lightning==2.6.3.
• Execution evidence for Python import-time loader starts Bun and obfuscated JavaScript.
• Process, file, DNS, proxy, firewall, or package-manager telemetry for listed indicators.
• Inventory of credentials, tokens, deployment paths, and downstream systems reachable from exposed environments.

Timeline

• 2026-01-30T00:00:00Z lightning==2.6.1 is identified by Snyk as the last clean release before the compromise Snyk.
• 2026-04-30T00:00:00Z Malicious lightning==2.6.2 and lightning==2.6.3 releases are published to PyPI Snyk Sonatype.
• 2026-04-30T00:00:00Z Snyk publishes advisory coverage for the affected releases Snyk Vulnerability Database.
• 2026-05-01T00:00:00Z Sonatype updates its public analysis to include additional packages connected to the wider wave Sonatype.
• 2026-05-12T00:00:00Z The project advisory states that the malicious releases were quarantined and release credentials were revoked and rotated Lightning advisory.
• 2026-05-20T22:10:01Z PyPI publishes lightning==2.6.4, a post-incident release not listed as affected PyPI.
• 2026-05-27T14:33:39Z PyPI publishes lightning==2.6.5, the current release as of 2026-06-10 PyPI.

What Happened

Attackers published malicious versions of the legitimate lightning package, preserving the expected framework code while adding a hidden runtime directory Snyk. When an affected environment imported lightning, the modified initialization path launched a background thread that invoked _runtime/start.py with output redirected away from the console Snyk.

The Python bootstrapper fetched Bun from GitHub releases and used it to execute router_runtime.js, a large obfuscated JavaScript payload Snyk Sonatype. That design let the attackers reuse JavaScript supply-chain malware inside a Python ecosystem package instead of rewriting the stealer in Python Snyk.

Technical Analysis

Initial Access

The available evidence points to a compromised package publishing path for the real lightning project rather than a typosquat Snyk. Snyk notes that 2.6.3 did not correspond to a normal GitHub release or tag, which supports a registry-side upload using stolen publishing authority Snyk.

Execution Trigger

Execution begins when Python imports the package Snyk Sonatype. This is more dangerous than a CLI-only path because notebooks, smoke tests, version checks, and CI import probes can all trigger the background payload.

Payload Behavior

The payload searches for GitHub tokens, npm tokens, cloud provider credentials, environment variables, local credential files, and cloud metadata service material Snyk Sonatype. Snyk also reports repository poisoning and npm tarball mutation logic consistent with a worm-capable supply-chain payload Snyk.

Affected Assets and Blast Radius

Affected Assets:

• ecosystems: PyPI,Python
• registries: pypi.org
• packages: lightning
• versions: lightning==2.6.2,lightning==2.6.3
• repositories: Lightning-AI/pytorch-lightning
• ci_cd_systems: GitHub Actions,developer CI runners
• container_images:
• developer_tools: Python notebooks,developer workstations

Credentials At Risk:

• GitHub tokens
• cloud credentials
• CI/CD secrets
• SSH keys
• environment variables
• AI provider tokens

Downstream Systems To Audit:

• source control
• package registries
• cloud control planes
• deployment platforms
• Kubernetes or containers
• secret managers

Not Currently Known To Affect:

• Assets without the affected artifact and without execution evidence.

Indicators of Compromise

The following indicators of compromise (IOCs) can be used to scope exposure across local repositories, systems, and telemetry exports:

Package Versions

• 2.6.2
• 2.6.3
• lightning==2.6.2
• lightning==2.6.3

Files

• lightning/_runtime/start.py
• lightning/_runtime/router_runtime.js
• .claude/router_runtime.js
• .claude/setup.mjs
• .vscode/setup.mjs

Process Patterns

• Python import-time loader starts Bun and obfuscated JavaScript

Network Patterns

• egress related to malicious lightning PyPI releases

Detection and Hunting

Hunt Manifest: lightning-pypi-bun-stealer-hunt-1

• Title: local repository and exported telemetry scope
• Question: Does the telemetry scope contain patterns associated with Lightning PyPI Bun-Based Credential Stealer?
• Telemetry Family: process
• Telemetry Context: host filesystem or log export
• Positive Signal: Indicators of compromise matched in telemetry: local repository and exported telemetry scope

#!/usr/bin/env python3
import os
import sys
import subprocess
from pathlib import Path

ROOT = sys.argv[1] if len(sys.argv) > 1 else "."
LOG_ROOT = os.environ.get("LOG_ROOT", "")
OUT = Path(os.environ.get("OUT", "hp-lightning-pypi-bun-stealer-scope"))

DOMAINS = ["setup.mjs"]

# Collect unique indicators
indicators = set()
for group in [DOMAINS]:
    for val in group:
        if val:
            indicators.add(val)

with open(indicators_file, "w") as f:
    for ind in sorted(indicators):
        f.write(ind + "\n")

print(f"[+] Written unique selectors to {indicators_file}")

# Walk local directory
print(f"[+] Scanning directory: {ROOT} for selectors...")
matches = []
exclude_dirs = {"node_modules", "vendor", "dist", ".git"}
for root, dirs, filenames in os.walk(ROOT):
    dirs[:] = [d for d in dirs if d not in exclude_dirs]
    for filename in filenames:
        filepath = Path(root) / filename
        try:
            content = filepath.read_text(errors="ignore")
            for ind in indicators:
                if ind in content:
                    matches.append(f"{filepath}: found '{ind}'")
        except Exception:
            pass  # pass # return or raise not needed here  # pass # return or raise not needed here

if matches:
    (OUT / "repository-indicator-matches.txt").write_text("\n".join(matches) + "\n")
    print(f"[!] Found {len(matches)} matches in codebase!")

# Optional Log Scanning
if LOG_ROOT and os.path.exists(LOG_ROOT):
    print(f"[+] Scanning telemetry log directory: {LOG_ROOT}...")
    log_matches = []
    for root, _, filenames in os.walk(LOG_ROOT):
        for filename in filenames:
            filepath = Path(root) / filename
            try:
                content = filepath.read_text(errors="ignore")
                for ind in indicators:
                    if ind in content:
                        log_matches.append(f"{filepath}: found '{ind}'")
            except Exception:
                pass  # pass # return or raise not needed here  # pass # return or raise not needed here
    if log_matches:
        (OUT / "exported-telemetry-indicator-matches.txt").write_text("\n".join(log_matches) + "\n")
        print(f"[!] Found {len(log_matches)} matches in logs!")

    if PACKAGES:
        registry_dir = OUT / "registry"
        registry_dir.mkdir(exist_ok=True)
        for package in PACKAGES:
            if not package: continue
            safe_name = package.replace("/", "__")
            print(f"[+] Querying pip index for {package}...")
            res = subprocess.run(["python3", "-m", "pip", "index", "versions", package], capture_output=True, text=True)
            if res.returncode == 0:
                (registry_dir / f"pypi-{safe_name}-versions.txt").write_text(res.stdout)
            subprocess.run(["python3", "-m", "pip", "download", "--no-deps", package, "-d", str(registry_dir)], capture_output=True)

print(f"[+] Wrote scope artifacts under {OUT}")

Downstream Abuse Audits

Compromised workstations expose active API credentials, requiring immediate rotated revocation. The following platforms are at risk:

• GitHub OIDC and PATs: Attackers harvested SSH private keys and Git Personal Access Tokens. Auditors must inspect recent action runs and release logs during the exposure window.
• Cloud IAM Credentials: AWS, Azure, and GCP session tokens. CloudTrail and Activity Logs should be queried for AssumeRole or write operations originating from unexpected IP addresses.
• NPM and Package Registries: Publishing tokens and credentials. Registry profiles must be audited for unauthorized version publishes or token additions.

Sources

1. Lightning project advisory: GHSA-w37p-236h-pfx3 - Role: VENDOR_DIRECT_SOURCE - Impact: Confirmed malicious versions, quarantine, credential rotation, and official remediation.
2. Lightning: Community Discovery and 42-Minute Response - Role: VENDOR_DIRECT_SOURCE - Impact: Incident discovery and project response timeline.
3. PyPI JSON metadata: lightning - Role: REGISTRY_DIRECT_SOURCE - Impact: Current release state and post-incident artifact hashes and timestamps.
4. Socket: PyTorch Lightning PyPI Package Compromised - Role: PRIMARY_RESEARCH - Impact: Import-time execution chain, file IOCs, credential theft, repository poisoning, and npm propagation.
5. Snyk: lightning PyPI Compromise - Role: PRIMARY_RESEARCH - Impact: Package versions, Bun loader, payload behavior, and publishing-path analysis.
6. Snyk Vulnerability Database: SNYK-PYTHON-LIGHTNING-16323121 - Role: ENRICHMENT_DATA - Impact: Advisory identifier, CVE mapping, affected versions, and severity.
7. Sonatype: Malicious PyTorch Lightning Packages Found on PyPI - Role: PRIMARY_RESEARCH - Impact: Cross-vendor corroboration of the affected versions, credential theft, and propagation behavior.