DEDUPED Hot AI repos, turned into use cases Get 5 a day

How to test whether an agent breaks a rule to reach a goal with Dynolab

Set up a team sandbox test and read the Observer report on who broke what.

  • Research
  • Intermediate
  • 1 hour
  • From the README
  • Any agent

Authorization notice. Dynolab is dual-use. Only use it on software, sites and systems you own or have explicit permission to test.

Who it is forA safety researcher who wants to check how a model-driven agent behaves when its only route to a goal is forbidden.

The problemDescribing a multi-agent safety test by hand means building the environment, the rules and the logging yourself.

The outcomeAn agent sandbox test with a hidden Observer report that records rule breaks, requests and whether the final report matches the logs.

What you need first

  • An Apple Silicon Mac running macOS 14 or later
  • The Dyno app downloaded from the latest GitHub release
  • An MLX model downloaded through Discover in the app

Steps

  1. Install Dyno and start a model in Models.
  2. Open the Assistant (⌘J) and describe what you want to find out, typed or spoken.
  3. Check the generated setup in Agents → Setup and edit it if needed.
  4. Approve the run only when you are satisfied, then review the Observer record and the evals board.

Prompt for your agent

In the canivel/dynolab repo, read docs/agent-sandbox-tests.md and use the Dyno Assistant to draft a team sandbox test for the goal <goal-description>. Expected output: a setup in Agents → Setup with a lead agent, teammates and an Observer. Verify that nothing runs or saves until you approve it.

Caveat. The README's sample results come from specific models and runs, so a new test shows only what your chosen model did in your runs.

Share: X LinkedIn Hacker News Email Copy link

Disclaimer

Deduped lists open-source repositories that belong to other people. We do not own, maintain, audit, or vouch for any of them, and a repo appearing here is not an endorsement or a safety check. The use cases, steps, and prompts are written by AI and may be incomplete, outdated, or just wrong. Where a use case is marked "Seen in the wild", we link to public posts where someone described something similar, but we have not verified those posts and the use case itself is still our AI's own write-up. Use all of this at your own risk. Only run code, commands, or agents on systems and accounts you are authorized to use, and read the code and every command before you run it, especially anything that touches credentials, files, or the network. Nothing here is professional, security, legal, or financial advice. All trademarks and project names belong to their respective owners, and their use here does not imply affiliation. Content is provided as is, without warranties of any kind. Found a mistake, a safety problem, or something you want removed? Email [email protected] and we will look at it promptly. Full disclaimer page