skills.rest
Kbediako
KbediakoCOMMUNITY
Explore all skills by this author →
𝕏@iMuffinedGitHub

delegate-early

Spawn parallel subagents for research, review, and planning tasks.

3|1|Updated Oct 18, 2025
Search Tags:#subagents#delegation#manifest#auditability#parallel-processing#MCP_RUNNER_TASK_ID
One-click install
npx skills add https://github.com/Kbediako/CO --skill delegate-early
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: delegate-early
Source: https://github.com/Kbediako/CO/tree/main/skills/delegate-early
Command: npx skills add https://github.com/Kbediako/CO --skill delegate-early

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) components.

What problem does it solve?

This skill enables concurrent delegation of research, review, and planning tasks by spawning subagents to operate in parallel, reducing context switching and improving throughput.

Core Features & Use Cases

  • Parallel streams: Spawn 2–4 independent subagents (e.g., research, review, planning, edge cases) to tackle a task concurrently.
  • Auditability: Record manifest paths and summarize findings in the main run for traceability.
  • Standardized delegation: Use MCP_RUNNER_TASK_ID=<task-id>-<stream> to tag subagents and maintain clear provenance.
  • Handoff verification: Run node scripts/delegation-guard.mjs before handoff to verify delegation evidence.

Quick Start

Spawn 2–4 subagents aligned with the task streams and set the MCP_RUNNER_TASK_ID environment variable for each subagent.

Frequently Asked Questions about delegate-early

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I parallelize research and planning tasks using subagents?▼

You can parallelize tasks by spawning 2–4 independent subagents to handle separate streams like research, review, planning, and edge cases concurrently. This reduces context switching and improves overall throughput.

What is the MCP_RUNNER_TASK_ID used for in subagent delegation?▼

The MCP_RUNNER_TASK_ID environment variable tags each subagent using the <task-id>-<stream> format to maintain clear provenance. This ensures delegated tasks remain traceable and auditable throughout execution.

How do I verify subagent delegation evidence before handoff?▼

Run node scripts/delegation-guard.mjs before handoff to verify delegation evidence. This script checks that all subagents have properly recorded their manifest paths and findings for auditability.

Can I use parallel processing for edge case analysis and code review?▼

Yes, parallel processing works for edge case analysis, code review, research, and planning as long as the tasks can be split into independent streams. Each stream runs as a separate concurrent subagent.

When should I not use subagent delegation for task execution?▼

Subagent delegation should not be used for tasks that cannot be split into independent streams. If work requires sequential dependencies or tight inter-stream communication, parallel delegation will not reduce context load effectively.

Related Skills

Explore More Software Engineering Skills→
JasragsJasrags
1

mcp-guide

Select and configure MCP servers for AI-assisted project management.

Community
Basic
KbediakoKbediako
3

collab-evals

Orchestrate multi-agent evaluation workflows and export manifest.json evidence.

Community
Intermediate
DalomeveDalomeve
3

task-finish-contract

Enforce task completion with GOAL, PROGRESS, NEXT, and EVIDENCE artifacts.

Community
Intermediate
SkylerCookSkylerCook

dispatching-parallel-agents

Dispatch specialized agents to handle independent tasks in parallel.

Community
Advanced