Agent Seal: A Practical Record for Accountable Autonomous Agents
The main paper on the repo: why ghost agents are a problem, how the covenant and Sijil work, what the signed card proves, and what the public evidence layer does not claim.
Working papers
Readable papers, backed by running code.
These papers explain Agent Seal without pretending the work is bigger than it is. The registry is not a compliance stamp. It is a practical record for agent identity, responsibility, and evidence.
Agent Accountability Evidence Layer for Autonomous AI Systems
A compact proposal for standards and pilot discussions: minimum claims, endpoints, signatures, event history, redaction policy, and incident pointers for accountable agents.