면책 조항

Agents are tools, not professionals. "Financial Analyst", "Legal Research Assistant", "Medical Research Assistant", "Reflection Assistant" and every other agent on this site are structured prompt templates. They simulate a style of analysis; they hold no license, carry no duty of care, and can be wrong. Their built-in disclaimers are part of the product and must not be removed when you use them.

No investment advice. Nothing here is a recommendation to buy or sell any security or asset, and no output should be read as one. Figures computed by finance-related agents depend entirely on inputs and stated assumptions. Agent of Me is not a registered investment adviser or broker-dealer. Verify everything against primary sources; decisions with real money deserve a licensed adviser.

No legal advice. Legal agents support research framing only. AI can misremember cases and statutes. Every citation must be independently verified before any reliance. No attorney-client relationship exists. Deadlines, filings, and signatures belong with licensed counsel in your jurisdiction.

No medical or mental-health advice. Research and reflection agents provide educational structure only, never diagnosis, treatment, or therapy. If you are in crisis or considering self-harm, contact local emergency services or a crisis hotline now. Health decisions belong with qualified clinicians.

No tax or accounting advice. Tax rules are jurisdiction-specific and change; agent output is research framing to discuss with a CPA or tax adviser.

Community-submitted agents. The Community tab carries agent configurations written by other users. We did not create them, we do not maintain them, and listing one is not an endorsement. Every submission is automatically scanned before it can appear, for embedded scripts and install commands, credential-harvesting and data-exfiltration instructions, links to executables, prompt-injection payloads and jailbreak content, and anything that fails is rejected while anything suspicious is held back. That scan is not a guarantee: it cannot catch everything and it does not inspect whatever a prompt points you to. Read any third-party prompt in full before using it, and never act on one that asks you to run a command, download a file, or hand over passwords, API keys or payment details. Report anything harmful via Contact and we will remove it.

Generated content can be wrong. All AI output, profiles, comparisons, test runs, answers, may contain errors or omissions. Review before use; you are responsible for what you send, sign, or act on.

Platform facts drift. Directory and guide details reflect their last-reviewed dates and can be outdated the day after; the official sites are authoritative.

See also: Terms of Use · Privacy Policy · Methodology.