Helping agents consider consequences
The Eevo Consequence-based Decision API is designed to help AI agents and humans make better decisions based on their consequences.
Eevo is designed to work with your agent: the agent provides a decision context, the options it has identified, and a description of their consequences. In practice, the agent calls the Eevo Consequence-based Decision API with that information, typically through the Eevo MCP server.
Eevo's harness then runs hundreds of individual calls to the underlying Eevo decision model to comparatively size the consequences, model their overlap, and compute expected value. The result is then returned to your agent, which uses the information to make a decision, or provides well-structured information for humans to make a decision.
The easiest way to integrate your agent with Eevo is to add the Eevo MCP server, which is available as a Connector in Claude's directory and as a ChatGPT plugin. It is also easy to set up manually. Instructions are on the Connecting your agent to Eevo page.
Suggested notes to add to AGENTS.md
Below are instructions we have found to be helpful to add to AGENTS.md.
* When writing code that calls Eevo's API, first read https://eevo.io/llms.txt, which links to every
page of the Eevo docs as Markdown and to the OpenAPI description of the API.
* When planning work and thinking of asking the user a question, and the correct answer depends on
how one weighs consequences, before asking the user, use Eevo through the relevant MCP tools to
compare options. If the question remains unclear, only then ask the user. Ideally show the user
the Eevo result, e.g. visualized in planning artifacts.
* When encountering unexpected situations not covered by your rules/instructions, use Eevo to help
decide what to do.
* Consider Eevo results informational. Remember that Eevo has only the context you provide to it. If
Eevo results are counterintuitive, consider what context, options, or consequences you may have
forgotten to include, and possibly call Eevo again. This process can help sharpen your thinking.