The Agent Skills Directory

[SAFE]: No malicious code or suspicious patterns were detected. The content consists of legitimate technical documentation for building voice-based AI agents.
[EXTERNAL_DOWNLOADS]: The skill references standard platform dependencies including @zavudev/functions and @zavudev/sdk, as well as scaffolding agents through the zavudev CLI. These are official resources provided by the vendor.
[COMMAND_EXECUTION]: Examples of platform-specific CLI commands are provided for deployment (npx zavudev deploy) and testing (npx zavudev agents test), which are necessary for the skill's documented purpose.
[INDIRECT_PROMPT_INJECTION]: The skill presents an attack surface where tools ingest external data (e.g., from api.example.com). Ingestion points: The check_availability tool in SKILL.md. Boundary markers: The skill explicitly instructs the agent in the prompt to never read URLs or code blocks aloud. Capability inventory: The skill uses fetch for network operations. Sanitization: None described. The risk is assessed as safe due to the presence of safety-oriented prompt instructions that limit the agent's output behavior when processing data.

voice-agent