JevMade field notes / Integration walkthrough

Let an AI assistant choose safe actions from a strict list

This guide explains how to force an AI assistant to pick from a strict list of safe actions. It shows how to check the risk of a chosen task before letting the software actually perform it.

Original by ComposioIntegrations

Listen to this guide

JevMade’s plain-English explanation

0:00 /

AI narration

Credits and license

“Route Composio tools with a typed Jev decision” by Composio. Read the original source.

This expanded guide is an AI-narrated adaptation of the source’s essential explanation, examples and caveats, not a word-for-word reading. The synthetic voice does not imitate the author or imply their endorsement.

License: MIT

MIT License

Copyright (c) 2025 Sampark Inc.

Permission is hereby granted, free of charge, to any person obtaining a copy
of this software and associated documentation files (the "Software"), to deal
in the Software without restriction, including without limitation the rights
to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
copies of the Software, and to permit persons to whom the Software is
furnished to do so, subject to the following conditions:

The above copyright notice and this permission notice shall be included in all
copies or substantial portions of the Software.

THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
SOFTWARE.

Our summary

This guide shows how to connect an AI assistant to outside services using a tool collection called Composio. It uses Jev, an AI tool that chooses from options rather than writing an answer. This helps prevent the software from inventing unsafe actions.

When a person asks for help, the software looks at the available tasks. It must pick one exact match from the list or decide to do nothing. If it picks a task, a safety check runs before the software actually carries out the action.

This setup is useful for developers who want to block harmful changes. The example only allows actions that read information. You need private access codes for Composio and TypeSafe to try this. It does not guarantee every available action is completely safe.

Key takeaways

Force the AI assistant to pick from a strict list instead of inventing its own actions.
Always allow the software to safely skip a task if it cannot find the right match.
Check the safety level of every chosen action before letting the software actually perform it.

More guides →See the experiments →