Refactor rule-expander agent to <3000 chars and enhance copilot-instructions with agentic workflow patterns

[Copilot]

The rule-expander.agent.md exceeded GitHub Copilot's 3000 character agent limit at 63,322 chars, making it unusable. The copilot-instructions.md lacked modern agentic workflow guidance.

Changes

rule-expander.agent.md (63,322 → 2,800 chars)

• Condensed to essential workflow: Research → Implementation → Testing → Quality → Delivery
• Retained core patterns: rule structure, discovery criteria, safety constraints
• Removed verbose multi-phase instructions and redundant examples
• Fixed priority guidance to match codebase reality: 50-100 (specific), 500-900 (high), 1000 (default), 2000+ (fuzzy)

copilot-instructions.md (+261 lines, no removals)

• Agentic workflows: Task scoping, acceptance criteria templates, incremental patterns
• Security: Dependency checks, code safety, rule evaluation constraints
• Quality standards: Rust error handling, type safety, 6-test minimum per rule
• Context: Architecture principles, common patterns, commit conventions
• File organization: Clear guidance on where changes belong

Example

Agent now follows GitHub Copilot best practices with concise frontmatter and focused guidance:

---
name: Rule Expander
description: Research, implement, and deliver command correction support for CLI tools
tools: ["*"]
---

## Workflow
1. Research & Discovery
2. Implementation (use is_app(), register in mod.rs)
3. Testing (min 6 tests: 2+ positive, 2+ negative, 1+ correction, 1+ edge)
4. Quality (cargo test && cargo clippy && cargo fmt)
5. Delivery (feature branch, conventional commits)

Both files now align with 2024-2025 GitHub Copilot standards for agentic development.

---

💡 You can make Copilot smarter by setting up custom instructions, customizing its development environment and configuring Model Context Protocol (MCP) servers. Learn more Copilot coding agent tips in the docs.