The Agent Skills Directory

[SAFE]: The skill acts as a legitimate documentation and configuration guide for the Zavu AI agent platform. All functionality described aligns with the stated purpose of building conversational agents.
[CREDENTIALS_UNSAFE]: No hardcoded credentials were detected. The provided code examples for multiple languages (TypeScript, Python, Go, Ruby, PHP) consistently demonstrate using environment variables or system environment lookups (e.g., process.env.PROVIDER_API_KEY, os.environ["PROVIDER_API_KEY"]) for API keys and webhook secrets.
[PROMPT_INJECTION]: The skill involves processing 'Inbound messages' through an LLM, which represents an indirect prompt injection surface. However, this is an inherent aspect of the conversational AI use case and no patterns suggesting intentional bypasses or malicious instructions were found.

ai-agent