Building with Cursor

X-Fidelity includes specialized AI tooling for Cursor IDE that significantly accelerates development. When working on the codebase in Cursor, you have access to slash commands, specialized subagents, and step-by-step skills.

Slash Commands

Use these commands in Cursor chat by typing /xfi-* for guided development workflows:

Development Workflow Commands

Command	Description
/xfi-review	Comprehensive code review using three expert subagents in parallel (code reviewer, security expert, testing expert)
/xfi-test	Run the test suite and automatically diagnose any failures
/xfi-build	Build the monorepo and diagnose build issues

Creation Commands

Command	Description
/xfi-create-plugin	Create a new analysis plugin with facts, operators, tests, and sample rules
/xfi-create-rule	Create a new analysis rule with proper structure and archetype integration
/xfi-create-archetype	Create a new archetype configuration for a project type

Debug and Fix Commands

Command	Description
/xfi-debug	Debug errors, exceptions, and unexpected behavior
/xfi-fix	Smart routing to the appropriate expert subagent based on problem type
/xfi-analyze	Troubleshoot X-Fidelity analysis issues (rules, plugins, output)

Release and Documentation Commands

Command	Description
/xfi-release	Prepare for release following the unified release workflow
/xfi-docs	Update documentation to stay in sync with code changes

Quality Assurance Commands

Command	Description
/xfi-consistency	Verify CLI and VSCode extension produce identical results
/xfi-security	Perform security review of code changes

Specialized Subagents

The AI has access to domain-specific expert subagents that provide deep expertise in different aspects of X-Fidelity:

Subagent	Expertise	Auto-triggers
xfi-build-expert	Turbo, yarn workspaces, esbuild, TypeScript compilation	Build failures, CI issues
xfi-testing-expert	Jest, coverage, integration tests, VSCode extension testing	Test failures, coverage gaps
xfi-plugin-expert	Plugin architecture, facts, operators, AST analysis with tree-sitter	Plugin development
xfi-vscode-expert	VSCode extension, webviews, tree views, diagnostics, packaging	Extension issues
xfi-rules-expert	json-rules-engine, archetypes, conditions, events, exemptions	Rule creation
xfi-security-expert	Path validation, directory traversal prevention, webhooks, secrets	Security reviews
xfi-debugger	Error analysis, log interpretation, StandardError patterns	Runtime errors
xfi-docs-expert	README, website docs (Docusaurus), CHANGELOG	Documentation updates
xfi-code-reviewer	Balanced code review, quality and security assessment	Before commits, PR reviews

Parallel Subagent Execution

Subagents can run in parallel when their domains are independent. For example, after implementing a feature:

Run in parallel:
├── xfi-code-reviewer (review the changes)
├── xfi-testing-expert (check test coverage)
└── xfi-docs-expert (check documentation needs)

Skills

Skills are step-by-step guides that are automatically loaded when relevant. They provide detailed workflows for common development tasks:

Skill	Purpose
xfi-create-plugin	Complete plugin creation workflow with templates
xfi-create-rule	Rule creation with fact/operator selection guidance
xfi-create-archetype	Archetype configuration with rule and exemption setup
xfi-release-workflow	Release process, conventional commits, version management
xfi-debug-analysis	Troubleshooting analysis issues systematically
xfi-consistency-testing	CLI-Extension parity verification workflow
xfi-documentation-update	Documentation sync between README and website
xfi-add-package	Adding new packages to the monorepo

File Locations

Resource	Location
Slash Commands	.cursor/commands/
Subagents	.cursor/agents/
Skills	.cursor/skills/
Workspace Rules	.cursor/rules/

Example Workflows

Creating a New Plugin

1. Use /xfi-create-plugin command
2. The AI reads the xfi-create-plugin skill for step-by-step guidance
3. The xfi-plugin-expert subagent provides architecture expertise
4. Complete plugin with facts, operators, tests, and sample rules is created

Debugging a Test Failure

1. Use /xfi-test command
2. Tests run via yarn test
3. If failures occur, xfi-testing-expert analyzes the issue
4. If it's a runtime error, xfi-debugger is also engaged
5. Fix is applied and verified

Preparing a Release

1. Use /xfi-release command
2. xfi-testing-expert and xfi-docs-expert run pre-checks in parallel
3. The xfi-release-workflow skill guides through conventional commits
4. Dry run is executed to preview the release

Tips for Effective Use

1. Let the AI auto-delegate: Subagents are triggered automatically based on context
2. Use slash commands for complex tasks: They provide structured workflows
3. Trust subagent expertise: Each subagent is a subject matter expert
4. Combine insights: Multi-subagent commands synthesize findings from all experts
5. Check the skills: Skills contain detailed checklists and templates