dialogue-audio
Pass
Audited by Gen Agent Trust Hub on Sep 8, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill directs users to install external components using 'npx skills add' from the 'belt-sh' and 'inference-sh' GitHub organizations. It also points to installation guidelines hosted at a raw GitHub URL ('https://raw.githubusercontent.com/inference-sh/skills/refs/heads/main/cli-install.md'). These sources are external to the agent platform's default trusted environment.
- [COMMAND_EXECUTION]: The skill is configured to use the 'Bash' tool to run the 'belt' CLI. This allows the skill to execute external commands on the system, which are used to authenticate and trigger remote audio generation tasks.
- [INDIRECT_PROMPT_INJECTION]: The skill processes user-supplied text for multi-speaker dialogue generation, which constitutes an ingestion surface for potential indirect instructions. * Ingestion points: Prompt fields within the 'belt app run' command examples (SKILL.md). * Boundary markers: Prompts are provided as raw text within JSON inputs without explicit 'ignore embedded instructions' markers. * Capability inventory: The skill has access to command execution via the 'Bash' tool. * Sanitization: No sanitization or validation of the input text is described before it is passed to the external CLI.
Audit Metadata