AI disclaimer
Review before use.
OpenAgentSource helps users inspect portable AI agent manifests. It does not make third-party agents safe by default.
AI Agents Can Be Wrong
AI agents and workflows may be inaccurate, incomplete, unsafe, biased, outdated, unavailable, insecure, or unsuitable for your intended use.
Agents May Access Tools or Data
Depending on their configuration, agents may take actions, access tools, call external services, or interact with data outside OpenAgentSource. Review the manifest, permissions, source links, dependencies, and install instructions before use.
Independent Verification Required
Users must independently verify agent behavior, outputs, permissions, source code, manifests, dependencies, package links, install instructions, and any downstream actions before relying on them.
Trust Signals Are Not Guarantees
Trust labels, risk levels, validation status, reviews, and creator profiles are informational signals only. They are not guarantees of safety, quality, legality, performance, or suitability.
No Professional Advice
OpenAgentSource does not provide legal, medical, financial, tax, investment, security, compliance, or other professional advice. Consult qualified professionals where appropriate.
No Control Over Third-Party Behavior
OpenAgentSource lists manifests and related metadata. We do not execute, host, operate, monitor, or control third-party agents, external endpoints, local runtimes, MCP servers, or AI app behavior.
High-Risk Uses
High-risk, sensitive, regulated, or consequential use cases require human review, additional testing, and appropriate safeguards before use.