The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The script interacts with public web services such as emailnator.com, mail.tm, and api.mail.tm to generate temporary email addresses and retrieve messages. These are well-known, legitimate services for disposable emails.
[COMMAND_EXECUTION]: The script uses os.system('cls||clear') for UI cleanup and subprocess.Popen to manage the lifecycle of Chrome browser instances with specific debugging ports. These are standard automation practices.
[DATA_EXFILTRATION]: The script saves generated session data (cookies and account info) to local files like cookies.txt, accounts.txt, and logs.txt. It does not send this data to any unauthorized remote server; it only interacts with the designated email provider APIs.
[PROMPT_INJECTION]: No evidence of instructions designed to bypass safety filters or override AI behavior was found.
[OBFUSCATION]: The code is written in plain Python with clear logic and variable names. No obfuscation techniques like Base64 encoding of commands or hidden characters were detected.
[REMOTE_CODE_EXECUTION]: The script does not download or execute arbitrary scripts or binaries from remote sources. It uses the selenium and undetected_chromedriver libraries for web automation within a controlled environment.
[PRIVILEGE_ESCALATION]: The script does not attempt to gain administrative privileges or modify sensitive system configurations.
[PERSISTENCE_MECHANISMS]: No attempts to establish persistent access or modify startup scripts were found.
[METADATA_POISONING]: The script's metadata (if any) and comments are consistent with its stated purpose as a utility tool.
[INDIRECT_PROMPT_INJECTION]: While the script processes data from external websites (email content), it does not feed this data back into a prompt in a way that could influence the AI's core instructions.
[DYNAMIC_CONTEXT_INJECTION]: No use of load-time shell execution patterns was found.

websh