Skip to content
GitHub Agentic Workflows

Custom Engines

Define custom GitHub Actions steps without AI interpretation for deterministic workflows.

engine: custom

Extended configuration:

engine:
id: custom
steps:
- name: Install dependencies
run: npm ci

All engines (Copilot, Claude, Codex, and Custom) support custom error pattern recognition for enhanced log validation. This allows you to define project-specific error formats that should be detected in agent logs.

engine:
id: copilot
error_patterns:
- pattern: "\\[(\\d{4}-\\d{2}-\\d{2}T\\d{2}:\\d{2}:\\d{2})\\]\\s+(ERROR):\\s+(.+)"
level_group: 2
message_group: 3
description: "Custom error format with timestamp"

Define multiple error patterns to catch different error formats:

engine:
id: claude
error_patterns:
- pattern: "PROJECT_ERROR:\\s+(.+)"
message_group: 1
description: "Project-specific error"
- pattern: "VALIDATION_FAILED:\\s+(.+)"
message_group: 1
description: "Validation error"
  • pattern (required): ECMAScript regular expression to match log lines
  • level_group (optional): Capture group index (1-based) containing error level (ERROR, WARNING, etc.). Use 0 to infer from pattern content.
  • message_group (optional): Capture group index (1-based) containing the error message. Use 0 to use the entire match.
  • description (optional): Human-readable description of what this pattern matches

Error patterns can be defined in shared workflows and imported:

shared/error-patterns.md:

---
engine:
error_patterns:
- pattern: "SHARED_ERROR:\\s+(.+)"
message_group: 1
description: "Shared error pattern"
---

Main workflow:

---
imports:
- ./shared/error-patterns.md
engine: copilot
---

Custom error patterns are merged with engine-specific built-in patterns during workflow compilation.

  • AI Engines - Complete guide to AI engines
  • Frontmatter - Complete configuration reference
  • Tools - Available tools and MCP servers