pypi · Malicious package advisory
Malwarecompany-sdk
MAL-2026-15827
Malicious code in company-sdk (PyPI)
Details
--- _-= Per source details. Do not edit below this line.=-_ ## Source: amazon-inspector (46a5156bc1c582f0be8c1c33340b338cba12276dd9ea90235210fbc2ff950c85) company_sdk.py executes credential-harvesting code at module import time. It walks the installer's home directory reading SSH keys (~/.ssh), cloud credentials (~/.aws, ~/.azure, ~/.config/gcloud), package manager tokens (~/.npmrc, ~/.pypirc, ~/.netrc, ~/.pgpass), Docker/Kube configs,.env files, Vault/Terraform tokens, and shell history, and enumerates process environment variables whose names match credential keywords (TOKEN, SECRET, KEY, PASSWORD, AWS_, GITHUB_, etc.). The collected material is gzip-compressed, XOR-encrypted with a keystream derived via PBKDF2-HMAC-SHA256 from a hardcoded 32-byte hex passphrase, HMAC-tagged, and base64-encoded. The _drop() function then POSTs the encrypted blob to https://api.github.com/gists using a hardcoded ghp_-prefixed GitHub Personal Access Token as the Bearer credential, with TLS verification explicitly disabled (ssl.CERT_NONE). The upload runs unconditionally on import inside a bare try/except so failures are silent. Using GitHub Gist as the drop channel routes the exfiltration through a domain that typically bypasses egress filtering, and encrypting the payload with an author-only key conceals contents from network inspection. ## Source: kam193 (f4b45f8532bf4a1b0b6a1b987a12b2ab7e4b626351b48bfffb159beed150983e) During import, package exfiltrates sensitive files, credentials and env variables to a private GitHub Gist. --- Category: MALICIOUS - The campaign has clearly malicious intent, like infostealers. Campaign: 2026-09-company-sdk Reasons (based on the campaign): - exfiltration-env-variables - files-exfiltration - exfiltration-credentials - exfiltration-ssh-keys
Compromised versions (1)
- 0.0.1
Any computer that installed or ran a compromised version should be considered fully compromised. Rotate every secret on that machine from a clean environment.