purpose-built-agent-teams

Builds a task-specific squad of specialist agents and runs it with Claude Code's agent-teams tools. Optionally sends read-only scout subagents across codebase areas the user approves, folds their findings into a proposed roster of seats with dependencies and acceptance criteria, executes seats in dependency order, and scores results in a verify-and-fix loop capped at three rounds — any criterion still failing after the third round halts the work and produces a failure report. Retires a team only after the user confirms, naming the team and its live task state. Use when a request splits into two or more separable roles or parallel workstreams and the team tools are present. Not for single-file edits, quick lookups, or strictly sequential work.

Version

v1.0.0

Size

17 KB

Updated

2026-10-12

Your client verifies the signature before reading. This explicit manual read is available on every plan. Free automatically loads five chosen skills; paid plans automatically load relevant skills across the library.