sapiom-agent-builder

This skill is documentation and templates for building autonomous agents using Sapiom. It does not contain obviously malicious code or instructions like remote download-and-execute, obfuscated payloads, or calls to attacker-controlled endpoints. The primary risks are operational: autonomous agents will hold and use credentials (SAPIOM_API_KEY, DB creds), depend on third-party packages, and make outbound network requests. Those properties are coherent with the stated purpose but increase the potential impact of misconfiguration or dependency compromise. Recommend auditing any templates and included dependencies before deployment, ensuring secrets are managed carefully (least privilege, rotate keys, avoid embedding in logs), pinning dependency versions, and performing a sample run and quality checks as the docs advise.