synthesize
Pass
Audited by Gen Agent Trust Hub on Mar 3, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill utilizes the
Bashtool to execute a local Python script located at~/.local/share/kokoro/tts_generate.py. This enables the agent to perform system-level operations within the user's environment. - [PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection via shell injection because user-provided text is interpolated directly into a bash command string.
- Ingestion points: Untrusted user input enters the agent context through the
[text to speak]argument defined inSKILL.md(via the--textparameter). - Boundary markers: Absent. The command examples demonstrate direct interpolation without delimiters or instructions to ignore embedded instructions.
- Capability inventory: The skill is granted access to the
Bashtool, which allows for full command execution on the host system. - Sanitization: Absent. No evidence of escaping, validation, or filtering of shell metacharacters (such as
;,|, or&) is present in the skill's instructions.
Audit Metadata