moa

Fail

Audited by Socket on Sep 18, 2026

4 alerts found:

Anomalyx2SecurityMalware
AnomalyLOW
SKILL.md

SUSPICIOUS: the skill's core purpose is coherent, but it routes user prompts and potentially project code/WIP through a dispatcher to multiple configurable external backends whose exact destinations and trust level are environment-dependent. The footprint is proportionate to a multi-agent consultation tool, yet data-flow opacity and reliance on third-party CLIs/provider configs create meaningful security risk even without clear malicious intent.

Confidence: 81%Severity: 58%
SecurityMEDIUM
tests/fake_claudex_cli.py

This module is primarily a configuration/flag validation fixture, but it contains a high-severity security weakness/indicator: when stdin includes the marker 'FAIL_WITH_SECRET', it directly prints CLIPROXY_API_KEY (or a fallback) to stderr and exits. There is no evidence of networking, command execution, or persistent malware in this snippet, but the stdin-triggered secret disclosure makes the package unsafe to run in environments where CLIPROXY_API_KEY is set and stdin can be influenced by an attacker.

Confidence: 82%Severity: 78%
AnomalyLOW
tests/fake_backend.py

No clear malware behavior is present (no network, no subprocess execution, no persistence, no credential theft). The main security concern is misuse potential: untrusted --project and untrusted JSON fields (notably name and mode) drive filesystem read/write/list/exists operations and reconnaissance-style outputs, and 'unsafe' mode can misrepresent capabilities. Treat this component as security-sensitive in integrations that pass untrusted inputs, because it can probe and modify files within the permissions of the running process.

Confidence: 68%Severity: 58%
MalwareHIGH
tests/fake_claude_cli.py

This module contains an explicit, attacker-triggerable secret disclosure mechanism: if untrusted stdin includes the substring "FAIL_WITH_SECRET", it reads ANTHROPIC_API_KEY from environment variables and prints it to stderr before exiting. Even though the script also performs various flag/sandbox validations and outputs a JSON fixture result, the secret-leak path overrides those guardrails, making this a serious supply-chain/testing fixture sabotage risk. No evidence of network exfiltration or command execution exists in the shown fragment.

Confidence: 90%Severity: 95%
Audit Metadata
Analyzed At
Sep 18, 2026, 02:11 PM
Package URL
pkg:socket/skills-sh/ele-yufo%2Funstuck%2Fmoa%2F@58d6c67d42d5109ca5599184e14fcf2aaa83037114e4316a38653af1bab5d0ab
Security Audit — socket — moa