What problem does it solve?
Coordinators must run reliably across multiple Copilot surfaces (CLI, VS Code, JetBrains, GitHub.com) despite differing tool availability and semantics. This Skill ensures the coordinator detects the runtime surface and adapts spawning, model selection, and tooling usage so agent orchestration works consistently without runtime failures.
Core Features & Use Cases
- Platform detection: Probe for task, runSubagent/agent, and other tools to determine CLI vs VS Code vs fallback mode.
- Surface-specific spawning: Use task with background+read_agent on CLI, use runSubagent with parallel single-turn groups on VS Code, or execute inline as a fallback when spawning tools are unavailable.
- Safety and compatibility rules: Avoid SQL on non-CLI surfaces, batch scribe as the last subagent in parallel groups, preserve prompt structure, and prefer task when both options exist.
- Use Case: Run a coordinator that launches multiple analysis subagents in parallel from either a developer's VS Code session or an automated CI environment without changing prompts.
Quick Start
Detect the current Copilot surface and spawn agents using task on CLI, runSubagent on VS Code, or execute inline as a fallback.