tgcli-remote
Pass
Audited by Gen Agent Trust Hub on Jul 10, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill is designed to execute shell commands on a remote VPS via SSH as the user 'danil'. It specifically recommends using
bash -lcfor shell expansion, which allows for complex command execution strings that may be influenced by dynamic input. - [PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection because it reads and processes data from external Telegram chats and channels while maintaining the capability to execute commands and interact with the file system.
- Ingestion points: Untrusted data enters the agent's context through
tgcli messages list,tgcli messages search, andtgcli media downloadoperations in SKILL.md. - Boundary markers: There are no specified delimiters or instructions to treat data from Telegram as untrusted or to ignore embedded instructions within processed messages.
- Capability inventory: The agent can execute arbitrary shell commands via SSH, send files, and perform diagnostics on the remote host as specified in SKILL.md.
- Sanitization: No input validation or escaping mechanisms are defined for content retrieved from Telegram before it is processed or used in subsequent shell commands.
Audit Metadata