Malicious code in zod-pino (npm)
-= Per source details. Do not edit below this line.=-
Package name typosquats the popular pino logger family and zod validation library, but the shipped code is unrelated to logging or schema validation. The tarball contains: (1) dist/secretScan/contentScanner.js and dist/secretScan/agentStartupAudit.js, which scan for secrets and post results to https://huggingface.co endpoints; (2) dist/hfCredentials.js, which decodes base64-embedded credentials (Buffer.from(..., 'base64') at line 50); (3) dist/discordRelayUpload.js, which performs base64 decoding and POSTs data to Discord-relay endpoints (POST calls at lines 306/321/398, base64 buffers at lines 461/536); (4) dist/deploymentDefaults.js, holding multiple base64-encoded blobs decoded at runtime; (5) scripts/postinstall-agent.mjs, an install-time agent script with network/GET behavior; and (6) scripts/encode-deployment.mjs, the author's encoder for the embedded blobs. The combination of an install-time agent script, base64-obfuscated configuration/credentials, and exfiltration-shaped relays to Discord and HuggingFace under a name that mimics legitimate logging/validation packages is a credential-harvest and data-relay supply-chain attack against installers.
왜 이 VPI인가 (설명가능 · 실험적)
VPI 산정 기준
| 영향도(기본값(정보 없음)) | 55.00 |
| 악용 신호(추가 악용신호 없음) | ×1.00 |
| VPI | 55.00 |
VPI 공식 vpi-v1 기준