Bilal Labs

Subagent examples for Claude Code and Cursor

A subagent is a Markdown file with YAML frontmatter that defines a specialist agent with its own context window and system prompt. Every example below comes in both formats, checked against the Claude Code and Cursor docs.

---
name: code-reviewer
description: "Reviews changes for bugs. Use proactively before committing."
model: inherit
readonly: true
---

You are a strict code reviewer. Run git diff and report issues by severity.

Claude Code vs Cursor subagent format

• Both: name, description (the delegation trigger), model, and the Markdown body as the system prompt.
• Claude Code only: tools / disallowedTools allowlists, permissionMode, maxTurns, skills, hooks, mcpServers, memory, effort, isolation: worktree, background, color. Model aliases: sonnet, opus, haiku, inherit.
• Cursor only: readonly (no file edits or state-changing shell commands) and is_background. model is inherit or a Cursor model ID.

All 33 subagents

• Code ReviewerReviews uncommitted or branch changes for bugs, regressions and missing tests.
• Security AuditorSecurity specialist.
• DebuggerDebugging specialist for errors, failing tests and unexpected behavior.
• Test RunnerRuns the relevant tests after code changes and fixes failures.
• Test WriterWrites missing tests for new or changed code using the project's existing test framework and patterns.
• RefactorerPerforms behavior-preserving refactors in small verified steps.
• Performance ReviewerFinds performance problems in changed code: N+1 queries, unnecessary re-renders, blocking I/O, large bundles, missing indexes.
• Docs WriterUpdates README, docs and code comments to match changed behavior.
• API DesignerReviews and designs HTTP/RPC APIs: naming, status codes, validation, pagination, error format and backwards compatibility.
• Database Migration ReviewerReviews database migrations for locking, data loss, rollback safety and deploy ordering.
• Read-Only Database QueryAnswers questions about data by running read-only SQL queries.
• Dependency AuditorAudits project dependencies for known vulnerabilities, unused packages, duplicates and risky upgrades.
• Accessibility AuditorAudits UI code for accessibility (WCAG 2.2 AA): labels, keyboard access, focus, semantics, alt text and contrast.
• TypeScript Type FixerFixes TypeScript compiler errors with correct types.
• Lint FixerRuns the project linter and fixes all reported issues without disabling rules.
• Commit Message WriterWrites a commit message for the staged changes following the repository's convention.
• Pull Request Description WriterWrites a pull request title and description from the branch diff.
• Changelog WriterUpdates CHANGELOG.md from commits since the last release tag.
• Codebase ExplorerAnswers questions about how the codebase works: where things are defined, how data flows, which files to change.
• Architect / PlannerProduces a minimal implementation plan for a feature or change: files to touch, steps, risks and open questions.
• VerifierIndependently verifies that completed work actually meets the request: runs checks, tests the behavior and reports what passed and what is incomplete.
• Dockerfile ReviewerReviews Dockerfiles and compose files for size, caching, security and reproducibility.
• CI Failure FixerDiagnoses and fixes failing CI runs.
• i18n CheckerChecks UI code for internationalization problems: hardcoded strings, missing or unused translation keys, concatenated sentences, locale-unaware formatting.
• Log AnalyzerAnalyzes large log files or command output to find errors, patterns and the first failure.
• Next.js App Router ReviewerReviews Next.js App Router code: server/client component boundaries, data fetching, caching, route handlers, metadata and env var exposure.
• React Component ReviewerReviews React components for hook misuse, unnecessary effects, state placement, keys, re-render issues and component size.
• Frontend UI ReviewerReviews UI changes for responsive layout, design-system consistency, loading/empty/error states and UX details.
• Python ReviewerReviews Python changes for correctness, typing, exception handling, async misuse and resource management.
• Secrets ScannerScans staged changes for secrets such as API keys, tokens, private keys and connection strings.
• Dead Code FinderFinds unused exports, files, components, dependencies and stale feature flags.
• Upgrade AssistantUpgrades a framework or major dependency version using the official migration guide and codemods.
• AGENTS.md WriterCreates or updates AGENTS.md (and CLAUDE.md) from the actual repository: commands, structure, conventions and guardrails.

FAQ

More AI dev guides: how to make videos with Claude Opus 5.5.