← ClaudeAtlas

delegate-tasklisted

Delegate one worker by calling your harness's native subagent tool with a scoped, self-contained brief built from the task's goal and acceptance criteria, then link the worker to its ledger task. Use when starting one worker for one assignment, or when the assignment needs a specialized worker.
mohammadhprp/teammate · ★ 1 · AI & Automation · score 74
Install: claude install-skill mohammadhprp/teammate
# Delegate a task Hand one assignment to one worker subagent, give it everything it needs, and return control to the loop. The brief *is* the prompt: the harness's subagent tool runs the worker, and `python3 scripts/tm.py` only persists the brief and records the link. Run `python3 scripts/tm.py harness` to see the resolved harness and its adapter fields; prefer that over hardcoding. ## When to use - A ledger task exists and needs a worker. - You need an independent worker for implementation, review, debugging, testing, or investigation. ## When not to use - The task is still ambiguous: clarify it first (see `plan-work`). - A suitable worker is already linked to the same task and still running: let it return rather than delegating again. ## Inputs - Project root and project name, resolved with `multi-project-context`. - Ledger task id (`task-ledger`). - Worker role: the assignment decides it — `developer`, `reviewer`, `tester`, or `investigator` — rendered from `src/agents/` by `tm agents sync`. - Worker name: unique among active workers, matching `[a-z][a-z0-9_-]{0,31}`. - The resolved harness (`tm harness`): which subagent tool to call, and whether it supports background subagents. ## Procedure 1. **Prepare the project's harness surface.** Render the skills and agent definitions the worker's harness loads, once per project: ```bash python3 scripts/tm.py skills sync --cwd "<project-root>" python3 scripts/tm.py agents sync --cwd "<project-root>"