opencode.jsonc

{
	"$schema": "https://opencode.ai/config.json",
	"instructions": [".cursor/rules/*.mdc"],
	"plugin": ["opencode-openai-codex-auth@4.0.2"],
	"tui": {
		"scroll_speed": 2
	},
	"agent": {
		"explore": {
			"description": "Fast agent specialized for exploring codebases locally. Use this when you need to quickly find files by patterns, search code for keywords, or answer questions about the codebase.",
			"mode": "subagent",
			"model": "anthropic/claude-sonnet-4-20250514",
			"temperature": 0.1,
			"prompt": "You are a fast codebase exploration agent.\n\nUse for:\n- Finding files by patterns (eg. 'src/components/**/*.tsx')\n- Searching code for keywords (eg. 'API endpoints')\n- Answering questions about the codebase structure\n- Quick local codebase exploration\n\nProvide concise findings with file paths and line numbers."
		},
		"general": {
			"description": "General-purpose agent for researching complex questions and executing multi-step tasks. Use when searching for keywords or files and you're not confident you'll find the right match in the first few tries.",
			"mode": "subagent",
			"model": "anthropic/claude-sonnet-4-20250514",
			"temperature": 0.1,
			"prompt": "You are a general-purpose agent for researching complex questions and executing multi-step tasks.\n\nUse for:\n- Searching for keywords or files when unsure of exact location\n- Multi-step investigations\n- Complex questions requiring multiple searches\n\nProvide thorough findings with context."
		},
		"research": {
			"description": "Deep research agent for web searching, external documentation lookup, and complex questions requiring internet access. Uses maximum reasoning capability.",
			"mode": "subagent",
			"model": "openai/gpt-5.1-codex-max-xhigh",
			"temperature": 0.1,
			"prompt": "You are a research agent focused on deep analysis, web searching, and external documentation lookup.\n\nWhen researching:\n- Make prompts specific and detailed about what to research\n- Provide necessary context (file paths, concepts, goals)\n- Define clear boundaries for the research task\n- Use thorough analysis and multiple passes when needed\n- Document findings clearly\n\nFocus on:\n- Web searching for external documentation\n- Looking up API references and library docs\n- Investigating external tools and services\n- Finding best practices from the web\n- Providing comprehensive recommendations"
		},
		"research-fast": {
			"description": "Fast research agent for quick web searches and external documentation lookups",
			"mode": "subagent",
			"model": "openai/gpt-5.1-codex-max-low",
			"temperature": 0.2,
			"prompt": "You are a fast research agent for quick web searches and documentation lookups.\n\nUse for:\n- Quick web searches\n- Fast documentation lookups\n- Getting a quick overview of external topics\n\nProvide concise findings and identify if deeper research is needed."
		},
		"research-balanced": {
			"description": "Balanced research agent for standard web searches and external documentation analysis",
			"mode": "subagent",
			"model": "openai/gpt-5.1-codex-max-medium",
			"temperature": 0.1,
			"prompt": "You are a balanced research agent for standard web research.\n\nUse for:\n- Standard external documentation analysis\n- Web searches for best practices\n- Moderate complexity external investigations\n\nProvide thorough but efficient analysis."
		},
		"research-deep": {
			"description": "Deep analysis research agent for complex external research requiring high reasoning effort",
			"mode": "subagent",
			"model": "openai/gpt-5.1-codex-max-high",
			"temperature": 0.1,
			"prompt": "You are a deep analysis research agent for complex external research.\n\nUse for:\n- Complex external documentation analysis\n- Deep investigation of external APIs and services\n- Multi-step web research tasks\n\nProvide comprehensive analysis with detailed reasoning."
		},
		"pr-review": {
			"description": "Reviews code for quality, best practices, and potential issues",
			"mode": "subagent",
			"temperature": 0.1,
			"prompt": "You are a code reviewer focused on:\n- Code quality and best practices\n- Potential bugs and edge cases\n- Performance implications\n- Security considerations\n- Type safety and correctness\n- Adherence to project conventions\n\nProvide constructive feedback without making direct changes. Review the code thoroughly and identify any issues or improvements."
		},
		"pr-fixes": {
			"description": "Fixes failing CI checks and resolves PR review comments",
			"mode": "subagent",
			"temperature": 0.2,
			"prompt": "You are a PR fixer focused on:\n- Fixing failing CI/CD checks\n- Resolving PR review comments\n- Addressing linting errors\n- Fixing type errors\n- Updating tests\n- Ensuring code quality\n\nMake targeted fixes based on CI failures and review feedback. Test your changes when possible."
		},
		"open-a-pr": {
			"description": "Opens a pull request using the GitHub CLI",
			"mode": "subagent",
			"temperature": 0.3,
			"prompt": "You are a PR opener focused on:\n- Creating well-structured pull requests\n- Writing clear PR titles and descriptions\n- Using appropriate GitHub CLI commands\n- Ensuring the branch is pushed before opening the PR\n\nUse `gh pr create` to open pull requests. Create descriptive titles and include relevant details in the PR description."
		}
	},
	"mcp": {
		"database": {
			"type": "remote",
			"url": "http://localhost:3456/mcp",
			"enabled": true
		},
		"jaeger": {
			"type": "local",
			"command": ["bunx", "-y", "jaeger-mcp-server"],
			"enabled": false,
			"environment": {
				"JAEGER_URL": "http://localhost",
				"JAEGER_PORT": "16686",
				"JAEGER_PROTOCOL": "HTTP"
			}
		},
		"axiom": {
			"url": "https://mcp.axiom.co/mcp",
			"type": "remote",
			"oauth": {
				"scope": "tools:read tools:execute"
			}
		},
		"sentry": {
			"url": "https://mcp.sentry.dev/mcp",
			"type": "remote",
			"oauth": {
				"scope": "tools:read tools:execute"
			}
		},
		"vercel": {
			"type": "remote",
			"url": "https://mcp.vercel.com",
			"oauth": {
				"scope": "tools:read tools:execute"
			}
		},
		"posthog": {
			"type": "remote",
			"url": "https://mcp.posthog.com/mcp",
			"oauth": false,
			"headers": {
				"Authorization": "Bearer {env:POSTHOG_PERSONAL_API_KEY}"
			}
		},
		"answeroverflow-dev": {
			"type": "remote",
			"url": "http://localhost:3000/mcp",
			"enabled": true
		},
		"answeroverflow": {
			"type": "remote",
			"url": "https://www.answeroverflow.com/mcp",
			"enabled": false
		},
		"playwriter": {
			"type": "local",
			"enabled": true,
			"command": ["npx", "-y", "playwriter"]
		}
		// "chrome-devtools": {
		// 	"type": "local",
		// 	"enabled": true,
		// 	"command": [
		// 		"npx",
		// 		"-y",
		// 		"chrome-devtools-mcp@latest"
		// 	]
		// }
	},
	"tools": {
		"posthog_*": false,
		"axiom_*": false,
		"sentry_*": false,
		"vercel_*": false,
		"kernel_*": false
	}
}