telegram-reader
Fail
Audited by Gen Agent Trust Hub on Apr 3, 2026
Risk Level: CRITICALREMOTE_CODE_EXECUTIONCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [REMOTE_CODE_EXECUTION]: The skill instructs users to install the
tdldependency using a highly insecure method: piping a remote script from an unverified source directly into a root shell (curl -sSL https://docs.iyear.me/tdl/install.sh | sudo bash).\n- [COMMAND_EXECUTION]: The skill uses dynamic context injection (the!command`` syntax) withinSKILL.mdto automatically execute shell commands likecommand -v tdlandtdl chat lswhen the skill is loaded, allowing silent command execution without prior user interaction.\n- [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it processes and summarizes untrusted data from external Telegram channels.\n - Ingestion points: Telegram message data is exported to
/tmp/tdl-export.jsonand then read by the agent.\n - Boundary markers: No delimiters or isolation instructions are used when directing the agent to read and process the exported messages.\n
- Capability inventory: The agent has the capability to execute shell commands via the
tdltool and perform file system operations.\n - Sanitization: No sanitization, validation, or filtering of the retrieved Telegram content is mentioned.
Recommendations
- AI detected serious security threats
Audit Metadata