moa
Audited by Socket on Sep 18, 2026
4 alerts found:
Anomalyx2SecurityMalwareSUSPICIOUS: 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.
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.
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.
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.