engineer-copilot-agent

Validate and optimize GitHub Copilot custom-agent contracts with discovery, routing, and minimization scripts.

Updated Mar 25, 2026
One-click install
npx skills add https://github.com/Tyler-R-Kendrick/copilot-auto-training --skill engineer-copilot-agent-tyler-r-kendrick
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: engineer-copilot-agent
Source: https://github.com/Tyler-R-Kendrick/copilot-auto-training/tree/main/skills/engineer-copilot-agent
Command: npx skills add https://github.com/Tyler-R-Kendrick/copilot-auto-training --skill engineer-copilot-agent-tyler-r-kendrick

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires pyyaml, and includes scripts (resource) and references (resource) and assets (resource) components.

What problem does it solve? GitHub Copilot custom agents stored as .agent.md files often drift out of sync with the real workspace: they reference stale tool names, invent handoff targets, and grow bloated with inlined standards and mechanical checks. This Skill audits and rewrites those contracts so triggering, routing, handoffs, and prompt size stay correct and lean. ## Core Features & Use Cases - Runtime surface discovery: Scan .github/agents/*.agent.md and skills/*/SKILL.md to enumerate real agents, skills, tools, and handoff pairs before editing any contract. - Contract validation: Check frontmatter fields, kebab-case names, trigger phrasing, and flag declared tools or handoff targets that do not exist in the discovered surface. - Body analysis and minimization: Detect oversized sections, deterministic instructions that belong in scripts, and stale backticked names, then apply a recursive loop that moves standards into references and checks into scripts. - Use Case: Your custom agent keeps telling users to call tools that no longer exist and references a retired handoff target. Run the discovery and validation scripts, correct the routing against the live inventory, and add eval coverage so invented routing fails loudly in the future. ## Quick Start Audit the custom agent at .github/agents/my-agent.agent.md against this repository's real agents and skills, fix stale routing, and trim the prompt body.

Frequently Asked Questions about engineer-copilot-agent

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

FAQPage Schema
How do I validate a GitHub Copilot custom agent file?

Run validate_agent.py with the path to the .agent.md file and the repository root. It checks required frontmatter fields, kebab-case naming, trigger phrasing in the description, and whether declared tools, child agents, and handoff targets exist in the discovered workspace surface.

How do I reduce the size of a bloated .agent.md prompt?

Use the recursive minimization loop: draft a lean routing layer, move standards and examples into references, move mechanical checks into scripts, then re-read and trim again. Stop when the body is mostly routing and ownership guidance.

Why does my Copilot custom agent reference tools that do not exist?

Agent contracts drift when the repository's tools, skills, or agents are renamed or removed. Run discover_runtime_surface.py to enumerate the real agents, skills, and tool surfaces, then reconcile the contract against the live session inventory before editing.

What Python version and dependencies are required to run the scripts?

The scripts require Python 3.11 or later and the PyYAML package for parsing frontmatter. They work in any repository that stores Copilot custom agents as .agent.md files alongside a skills directory containing SKILL.md files.

When should I not use this skill for prompt optimization?

Do not use it for prompt-only or skill-only work unless the target artifact is a Copilot custom agent contract or you explicitly want to design one. General prompt engineering or skill authoring should use dedicated prompt or skill engineering workflows instead.