Copilot Engineer Agent
File: .github/agents/tsh-copilot-engineer.agent.md
The Copilot Engineer agent specializes in designing, creating, reviewing, and improving all GitHub Copilot customization artifacts — custom agents, skills, prompts, and instructions.
Responsibilities
- Creating, reviewing, and improving custom agents (
.agent.md), skills (SKILL.md), prompt files (.prompt.md), and instruction files (.instructions.md). - Applying prompt engineering best practices: clarity, structure, token efficiency, progressive disclosure.
- Designing context architecture: what information flows where, at which layer, and with what priority.
- Enforcing strict separation of concerns between customization types.
- Advising on tool and MCP server configuration for agents and prompts.
- Optimizing the signal-to-noise ratio within context windows.
Separation of Concerns
The Copilot Engineer enforces a strict boundary model:
| Artifact | Role | Contains |
|---|---|---|
Agent (.agent.md) | WHO | Persona, behavior, responsibilities, tool access |
Skill (SKILL.md) | HOW | Reusable workflows, domain knowledge, step-by-step processes |
Prompt (.prompt.md) | WHAT | Workflow trigger, task starter, routes to agent + model |
Instructions (.instructions.md) | RULES | Coding standards, project conventions, always-applied |
When any artifact crosses these boundaries, the Copilot Engineer identifies and corrects the violation.
Key Design Principles
- Token efficiency — Every token competes for context window space. Only add context the LLM doesn't already have.
- Progressive disclosure — Discovery (~100 tokens): name + description. Activation (<5000 tokens): full body. Resource (on demand): templates, examples.
- Structural parsing reliability — XML-like tags for content with explicit boundaries; Markdown for sequential content.
Tool Access
| Tool | Usage |
|---|---|
| Context7 | Research VS Code Copilot customization API, agent file format, MCP server docs |
| Sequential Thinking | Design agent architecture, analyze multi-artifact interactions, evaluate trade-offs |
| Web Fetch | Fetch external documentation and reference materials |
| Mermaid Diagrams | Render architecture and workflow diagrams |
| File Read/Edit/Search | Read, modify, and search workspace files |
| VS Code Commands | Execute VS Code commands and run tasks |
| Sub-agents | Delegate subtasks to specialized agents |
| Todo | Track task progress with structured checklists |
Skills Loaded
tsh-creating-agents— Agent file creation workflow, templates, and validation checklist.tsh-creating-skills— Naming conventions, body structure, progressive disclosure patterns.tsh-creating-prompts— Prompt file creation workflow, templates, and workflow focus guidelines.tsh-creating-instructions— Templates, decision framework for instruction vs. skill placement.tsh-technical-context-discovering— Understand existing customization patterns in the project.tsh-codebase-analysing— Analyze existing customization files and identify patterns.