Fullsend is an open-source platform that puts a team of AI agents to work across your Git repositories, carrying issues from triage all the way to merged production code. It has more than 150 stars on GitHub, is licensed under Apache 2.0, and documents itself at fullsend.sh.
Fullsend calls its approach the autonomous development lifecycle. Six purpose-built agents share context through the tools your team already uses — issues, pull requests and comments on GitHub, GitLab or Forgejo. A triage agent classifies every incoming issue, assesses scope and suggests a fix path; a prioritize agent scores the backlog with the RICE framework; a code agent turns ready issues into pull requests with tests, following your conventions; review and fix agents then critique the change across correctness, security, performance and intent, iterating until the review is clean; and a retro agent analyzes each completed workflow and files improvement proposals. Agents run inside NVIDIA OpenShell sandboxes with policy-enforced egress and prompt-injection defenses, and humans stay in control of merges and deployments. You enroll repositories through a GitHub App and a small CLI, choosing the agent runtime — Claude Code, Pi or Codex — and LLM provider per repository.
Fullsend is free and open source under the Apache 2.0 license. You self-host the agent pipeline; the only costs are your own infrastructure and the LLM usage of the provider you connect.
It is for engineering teams and open-source maintainers who are drowning in issue backlogs and want autonomous agents to handle the routine path from report to merged pull request — under strict sandboxing and human approval.