eigen-compute
Audited by Socket on Feb 20, 2026
1 alert found:
Malware[Skill Scanner] Instruction to copy/paste content into terminal detected All findings: [CRITICAL] command_injection: Instruction to copy/paste content into terminal detected (CI012) [AITech 9.1.4] [HIGH] supply_chain: URL with free hosting platform or high-abuse TLD detected (SC007) [AITech 9.1.4] This README + helper snippet appears to be legitimate documentation for a TEE deployment skill and the attestation code simply reads a local KMS public key and constructs a fingerprint. There is no direct evidence of malware or obfuscated/executable payloads. However the documentation contains risky operational guidance (encouraging exposing logs via HTTP as a workaround), instructs a global npm install (increasing supply-chain attack surface), and references an external verification endpoint without specifying trust/privacy details. Those factors make this skill moderately risky from a supply-chain and data-leak perspective even though no explicit malicious code is present. LLM verification: This SKILL.md is coherent with its stated purpose (TEE app deploy/attest/manage). I find no direct malicious code in the provided documentation or attestation snippet. The primary risks are supply-chain and operational: installing a global third-party CLI from npm without version pinning, reliance on a custom verifier domain, and documentation patterns that could encourage accidental exposure of sealed secrets via logs. Treat this as suspicious from a supply-chain perspective (not confirmed malw