Research
We study the three moments where an agent's trustworthiness can be tested, and map the rest of the field to decide where to look next. Every project and paper ships with its code and data.
Working tools, testbeds and methods. Filter by program.
All 7 projects
A protocol in which each claim one AI agent sends another carries evidence: records from an append-only Merkle log that the agent cannot write, signed by an attesting signer the agent cannot control.
Runs each AI agent job in its own micro-VM with no network card. Every way out is a door on the host, decided and signed by a Tracekit signer the agent cannot reach.
A flight recorder for AI agents. It records what an agent was asked, what it said and what it did, and makes any later edit to that record visible.
Finds which input actually made a team of AI agents act. It records what every model call saw, then replays the agents without each suspect to confirm the cause, without repeating any real side effects.
Transaction guards for AI agents that hold crypto wallets, whose safety promise holds at the instant the transaction executes.
A simulated firm run entirely by AI roles, made to reinvent itself through nine eras of technology history.
A map of where AI safety effort goes, used to decide where the lab should work.
Our published papers, in full, newest first. Each one explains the method, the results and what did not work.
When one AI agent tells another that “the tests passed” or “a person approved this refund”, the receiver has only the sender’s word.
AI agents now do real jobs: they install packages, call model APIs with their operators’ keys, and change code in repositories that deploy to production.
AI agents run shell commands, edit repositories, send email and move money with their operators’ credentials, and the record of what they did is usually a log that the agent, or whoever runs it, can edit.
AI agents that control wallets read attacker-reachable content, so they can be steered into proposing harmful transactions.
Multi-agent LLM systems are increasingly structured like organizations, with roles, critics and shared memory, yet they are evaluated on tasks that last minutes.
Collaborate
We co-author studies with researchers and teams. Send the question you want answered, or see how we work with companies.