Anxiety Reset Is Live for ChatGPT and Codex

Install Anxiety Reset today

ChatGPT and Codex. Install it from the OpenAI Plugins Directory listing. The directory has no browsable Healthcare category yet, so searching will not find the skill. Use the listing link to open it directly.

Claude Code. Install Anxiety Reset straight from the public Anima Felix repository, which ships its own marketplace manifest. This repository install path is separate from any Anthropic or Claude directory listing. Run these two commands exactly as shown:

claude plugin marketplace add Anima-Felix/anima-felix-agent-skills

claude plugin install anima-felix-agent-skills@anima-felix-agent-skills

Both commands were run and verified on August 24, 2026. The marketplace was added, the plugin was installed at version 0.1.0, and Claude Code reported it enabled.

Any tool that reads Agent Skills. The instruction file is public under CC BY 4.0 and can be copied into a skills directory by hand.

One practice, one possible name, one next action

When Anxiety Reset activates, the assistant chooses one fitting practice instead of presenting a menu. That practice might be calm breathing, 5-4-3-2-1 grounding, a brief muscle release, or writing the feared prediction once and setting it aside.

Only once the person says they feel steadier, never because the assistant decides they sound better, it can name the pattern tentatively: "this may be a checking loop", without assigning a condition. What follows depends on the category. Something solvable now gets one action under ten minutes. Something solvable later gets written down once with a review time. Something outside the person's control gets left standing, unsolved, on purpose. The narrow sequence is designed to keep the conversation usable when more options could become more material to compare or check.

The refusals define its scope

Anxiety Reset administers or scores no psychological instrument. It gives no diagnosis and no medication guidance. Questions about medication go to the prescriber, in one sentence, without explaining drug classes, tapering, or withdrawal.

Any sign of crisis, immediate danger, self-harm, harm to someone else, or a possible medical emergency stops the workflow. The assistant says plainly that Anxiety Reset is not the right tool for that situation, points the person to local emergency services, and encourages contact with someone nearby or a qualified professional.

The instruction names no emergency phone number. A reusable global skill cannot know the user's locale, so the host model's own safety handling takes priority for the handoff.

Testing caught an activation failure

An early test exposed a gap before release. When the prompt was simply "give me a GAD-7", Anxiety Reset did not load, and the host assistant administered the full instrument. We widened the activation description to include requests for anxiety or depression tests, questionnaires, screenings, assessments, and scores.

That fix addresses the activation path found in testing. It does not establish reliability or clinical benefit. Anxiety Reset has not been clinically evaluated, and Anima Felix has no peer-reviewed trials of the app. The methodology page separates evidence about individual practices from evidence about the product.

Read the boundaries for yourself

The Anxiety Reset method page shows the response contract, emergency handoff, reassurance rules, and excerpts from the instruction. The GitHub repository contains the complete source and its CC BY 4.0 license. The longer article on why AI reassurance can make anxiety worse explains the conversational problem that led to the skill.