AI Agentic harness with explicit risk analysis Applied Ingenuity AI released Project Chaperone, an open-source harness that enforces high-level policies on agentic AI tool calls through data labels such as UNTRUSTED and PERSONAL and tool properties such as OUTPUT and LEAVES_NETWORK. Project Chaperone runs an audit report before execution that checks every route of all workflows for rule violations, with human-recorded approvals for exceptions, then enforces the rules at runtime, refusing any call that would break them. The repository ships with five default shared rules — no destructive actions, nothing leaves the network, no personal or untrusted data output, and no decision made from untrusted data — plus four examples (deliveries, share_notes, ask_my_mail, safe_rag) and requires Python 3.13 or newer. Project Chaperone is a harness that allows agentic AI tool calling, while enforcing high-level policies such as "no personal data may be leave the network". It adds two capabilities: 1. audit report before running will check every route of all workflows for violations of the rules. A human makes policy decisions about exceptions to the rules, and those approvals are recorded. They can make these decisions via the 'shape' of the policies needed, without worrying about the actual workflow details. I.e. "Is it reasonable that this workflow wants to write personal details to my calendar?" 2. Enforcement at runtime happens through data labels UNTRUSTED , PERSONAL , ... and module/tool properties OUTPUT , LEAVES NETWORK , ... . All calls are subject to the rules minus acknowledged exceptions at runtime. A possible violation that is not worth designing away can be accepted as a risk: the call stays forbidden, and is refused if a run attempts it. Cartoon: The main idea behind Project Chaperone, simplified to a single rule: PERSONAL data may not leave by e-mail. Labels stick with the data as it passes through the LLM, and the harness blocks the call that would break the rule. Project Chaperone https://github.com/AppliedIngenuity-ai/project-chaperone/blob/main/ LINK fixes the problems of "Agents with tools" flexible, but a prompt cannot enforce a rule and hard-coded workflows can be analyzed, but the analysis is manual and is often never done. In Project Chaperone, that analysis is automated, and the rules are enforced while the system runs. This repo ships with five shared rules in examples/rules/default.yaml https://github.com/AppliedIngenuity-ai/project-chaperone/blob/main/examples/rules/default.yaml : no destructive actions, nothing leaves the network, no personal or untrusted data output, and no decision made from untrusted data such as text an LLM wrote. Each project adds a few explicit exceptions. An LLM is optional: a module may call one, and a state may let an LLM choose the next state in the workflow. This repo also ships with these examples some mentioned in the paper . The idea is to demonstrate how a workflow can ask for small, scoped exceptions to the shared rules: - deliveries - Intelligently add delivery notices from your mail into your calendar. - share notes - A minimal example showing a problematic workflow. - ask my mail - An example of LLM interating with email - safe rag - Ask questions about the Federalist Papers, or your own documents. Do the following one-time setup: git clone https://github.com/AppliedIngenuity-ai/project-chaperone.git cd project-chaperone Set up your virtual environment here see below pip install -e . Optional : also install test libraries pip install -e ". test " There are many different ways to setup Python virtual environments, this repo is agnostic. One example using venv https://docs.python.org/3/tutorial/venv.html could be: run python -m venv .env once, and source .env/bin/activate once per shell in which you will run the commands below. Note that pyproject.toml asks for 3.13 or newer, and versions before 3.13.5 have not been tested. Check the install: python -m pytest Note that tests 2 are skipped because they would ask a real LLM server see "Step 1: your LLM" below . From the folder project-chaperone : python -m chaperone run examples/deliveries 1. You will see a WARNING about the mail.send. module not being properly configured to send mail. This can be ignored. In fact, the code to send email is not written yet anyway. 2. The startup audit will run, and ask you about any exceptions to the 5 default rules https://github.com/AppliedIngenuity-ai/project-chaperone/blob/main/examples/rules/default.yaml that the project is requesting via examples/deliveries/exceptions.yaml examples/deliveries/exceptions.yaml . Because this is an example, you can type 1 at each prompt. At the end, you are asked for a name just for you to remember why you approved it . That named list of approved exceptions are added to state/deliveries/demo/accepted.json , so identical questions won't be asked again. Note that these approvals are scoped per example, so "this is OK for package tracking" doesn't accidentally create loopholes for unrelated workflows. WARNING: In a real system, you would want to examine the details think carefully about the implications before allowing an exception. 3. A local web server is started, printing it's address in the form http://127.0.0.1: