team-tech-debt

Technical debt remediation across single or multi-repo codebases is automated by orchestrating scan, assess, plan, fix, and validate phases using a coordinator and team-worker agent architecture. This systematically reduces risk and maintenance cost with minimal manual coordination.

Scanning a codebase to generate a prioritized remediation plan involves running a coordinator agent that delegates tasks to role-based workers like scanners and assessors. These agents evaluate code quality, assess severity, and produce a phased remediation plan.

Automated code fixes run in isolated session git worktrees where executor workers implement changes and validator workers run checks. This supports a fix-verify loop with rollback and accept policies to validate changes safely before opening a pull request.

Multi-agent orchestration works for parallelizing code quality tasks by assigning distinct roles like scanner, assessor, planner, executor, and validator to different team-worker agents. This architecture maintains clear ownership and coordinates the workflow via TaskCreate and TaskUpdate APIs.

Tracking the lifecycle of automated tech-debt fixes is best handled through a coordinator architecture using TaskCreate and TaskUpdate APIs. This manages the end-to-end governance from initial scan to final validation, ensuring every remediation step is logged and visible.

Automating tech-debt discovery and remediation requires repository read and write access along with agent and task APIs. These permissions allow the multi-agent system to create worktrees, execute fixes, and update remediation statuses across the codebase.