The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses a Bash block to invoke uv run for executing a Python script (send.py). This involves running shell commands that utilize environment variables such as CLAUDE_PLUGIN_ROOT to locate the script.
[NO_CODE]: The core functional logic for interacting with the Telegram API is contained in scripts/send.py. This external file is referenced but not provided for analysis, preventing a full audit of its internal security behavior.
[PROMPT_INJECTION]: The skill interpolates untrusted user data—specifically the title and about description parameters—directly into a Bash command block. Ingestion points: Untrusted data enters the agent context through the parameters in the bash usage example. Boundary markers: No delimiters or 'ignore' instructions are present to distinguish user data from the command structure. Capability inventory: The skill allows arbitrary command execution via the Bash tool to launch the script. Sanitization: No escaping or validation is specified for the interpolated strings, creating a surface for indirect prompt injection or command injection if user inputs contain shell metacharacters like backticks or dollar signs.

create-group