The Agent Skills Directory

[COMMAND_EXECUTION] (HIGH): Unsafe shell interpolation in the Windows fallback mechanism.\n
Evidence: In scripts/clipboard_manager.py, the set_clipboard function uses an f-string to construct a PowerShell command: f'Set-Clipboard -Value "{text}"'. This string is then executed using subprocess.run.\n
Risk: If a user or an automated process provides input containing double quotes and command separators (e.g., \"; calc; #), the shell will terminate the intended command and execute the injected payload.\n- [DATA_EXFILTRATION] (LOW): Arbitrary file system access via command-line arguments.\n
Evidence: The script implements --file for reading and --output for writing any file path accessible to the current user via the open() function.\n
Risk: While this is intended for clipboard management, it allows an agent to be manipulated into reading sensitive files (e.g., SSH keys, credentials) or writing malicious files to sensitive locations.\n- [PROMPT_INJECTION] (LOW): Indirect prompt injection surface through system clipboard ingestion.\n
Ingestion points: Untrusted data enters the agent context via the get_clipboard() function in scripts/clipboard_manager.py.\n
Boundary markers: Absent; the content is processed as raw text without delimiters or instructions to ignore embedded commands.\n
Capability inventory: The skill possesses command execution (subprocess.run) and file system manipulation (open) capabilities.\n
Sanitization: Absent; the script does not sanitize, escape, or validate the clipboard content or file paths before processing.

clipboard-manager