Researchers introduce grite, an open-source coordination substrate that enables autonomous coding agents to track shared work through git-based event logs, reducing duplicate efforts from 78% to 0% while tripling useful throughput. The system addresses a critical gap in multi-agent collaboration that traditional pull-request metrics cannot capture, revealing previously invisible failure modes like conflicting edits and lock starvation.
The emergence of autonomous coding agents capable of opening millions of pull requests has created a coordination paradox: while agents produce code faster, their acceptance rates lag significantly behind human developers. This paper identifies the root cause as a pre-PR coordination problem—agents lack effective mechanisms to communicate about ongoing work, leading to wasted effort and conflicts invisible to standard repository telemetry.
Grite solves this through an elegant design: a decentralized coordination layer that leverages git itself as a distributed ledger. By storing cryptographically signed, append-only event logs directly in repositories, the system provides agents with shared awareness of task claims and progress without requiring centralized infrastructure. This approach proves remarkably effective, reducing redundant work from 78% to near-zero while maintaining state consistency across all agent copies.
The research reveals concrete failure modes previously undetectable in pull-request history—race conditions, lock starvation, and conflicting edits become automatically recoverable with full provenance. For the AI development community, this represents a maturation moment: autonomous agents are shifting from isolated tools to collaborative entities requiring coordination primitives similar to human teams.
The implications extend beyond immediate productivity gains. As agent swarms become more prevalent in software development, coordination infrastructure becomes critical infrastructure. Grite's design—leveraging existing git infrastructure rather than demanding new centralized systems—positions it as a practical foundation for scaling agent deployment. The released dataset and mining toolkit enable further research into agent interaction patterns, potentially informing better incentive structures and trust mechanisms for human-AI collaborative development.
- →Grite reduces duplicate agent work from 78% to 0% while tripling useful throughput using a decentralized git-based coordination system.
- →The coordination problem between autonomous agents occurs before pull requests and requires new telemetry mechanisms beyond standard repository metrics.
- →Cryptographically signed event logs stored in git provide consistent state across distributed agents without centralized servers or silent write failures.
- →Mining coordination logs reveals failure modes invisible in pull-request history, including race conditions, lock starvation, and redundant task discovery.
- →The open-source release of dataset and toolkit enables broader research into multi-agent coordination patterns in software development.