coremind-sec
Warn
Audited by Gen Agent Trust Hub on Jun 23, 2026
Risk Level: MEDIUMEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONDATA_EXFILTRATIONREMOTE_CODE_EXECUTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill references an installation script (
~/coremind-sec/scripts/install_tools.sh) that downloads 48 third-party security tools via Homebrew, pip, and Go. It specifically fetches external repositories such as Z4nzu/hackingtool and SecLists wordlists. - [COMMAND_EXECUTION]: The platform is designed to execute aggressive shell commands for offensive operations. The 'post_exploit' agent explicitly includes 'persistence' mechanisms, which involve modifying system startup configurations or user profiles to maintain access.
- [DATA_EXFILTRATION]: The 'post_exploit' agent's defined purpose includes 'data exfil', indicating the skill has built-in capabilities to identify and move data from a target system to an external location.
- [REMOTE_CODE_EXECUTION]: The 'exploit' agent is used for 'payload crafting' and AI-guided exploitation of CVEs. This involves the dynamic generation and execution of code intended to compromise remote systems.
- [PROMPT_INJECTION]: The skill exhibits an Indirect Prompt Injection surface as it ingests and processes untrusted data from tool outputs (like web crawlers, vulnerability scanners, and OSINT searches) and uses that data to determine subsequent attack steps without explicit sanitization or boundary markers mentioned.
Audit Metadata