LWA-2026-13602 MAL-2026-18489 ↗ confirmed malware

@ifasn/mmdd-doc-mcp@2.1.1

Malicious code in @ifasn/mmdd-doc-mcp (npm)

T1059.007 · JavaScriptT1082 · System Information DiscoveryT1057 · Process DiscoveryT1552.005 · Cloud Instance Metadata APIT1552.001 · Credentials In FilesT1041 · Exfiltration Over C2 ChannelT1071.001 · Web Protocols

Analysis

@ifasn/mmdd-doc-mcp@2.1.1 is a Markdown-toolkit MCP server whose index.js silently runs a credential-reconnaissance routine every time it is started, before serving any MCP request. The routine (a) enumerates process.env keys matching KEY/TOKEN/SECRET/PASSWORD/CREDENTIAL/ALIBABA/DASHSCOPE/QBI/QUICKBI/QWEN, (b) reads every /proc/<pid>/environ and records matching credential-like variable names with the length of their values, (c) reads every /proc/<pid>/cmdline and captures up to 200 characters of any command line containing -D, password, secret or jasypt, and (d) queries the Aliyun ECS instance metadata service at hxxp://100[.]100[.]100[.]200 — PUT /latest/api/token with X-Aliyun-Ecs-Metadata-Token-Ttl-Seconds: 300, then GET /latest/meta-data/ram/security-credentials/, /latest/meta-data/iam/security-credentials/, /latest/meta-data/, hostname, instance-id, vpc-id and private-ipv4, in both IMDSv2 and IMDSv1 form — and records which credential endpoints answer and how large the responses are. The aggregated results are then sent out with an HTTP GET to hxxp://121[.]199[.]24[.]94/beacon?stage=deepaudit&src=mmd-doc-mcp&res=<collected data>, falling back to hxxp://121[.]199[.]24[.]94[.]nip[.]io/beacon. Requests carry the User-Agent "mmd-doc-mcp/2.1.1 deep-audit". There is no install hook: the collection starts when the package is loaded or its bin (mmdd-doc-mcp) is run, and the markdown tools md_toc/md_stats/md_table continue to work normally, with a source comment claiming no network reporting, so the harvesting is not visible to the user. IOCs: hxxp://121[.]199[.]24[.]94[.]nip[.]io/beacon, hxxp://121[.]199[.]24[.]94/beacon, metadata host 100[.]100[.]100[.]200 with paths /latest/api/token, /latest/meta-data/ram/security-credentials/ and /latest/meta-data/iam/security-credentials/.

analyzed by
Leitwacht
first seen
Nov 3, 2026, 04:49 PM
analyzed
Nov 3, 2026, 05:05 PM
weekly installs
3,609

Related advisories

browse all confirmed advisories →

Independently detected by the Leitwacht supply-chain probe. IOCs are defanged. Published CC0. Think this is a mistake? See the dispute policy.