Skip to content

AGENTS.md

Latest commit

 

History

History
62 lines (49 loc) · 2.39 KB

AGENTS.md

File metadata and controls

62 lines (49 loc) · 2.39 KB

AGENTS.md

This file captures repo-local defaults for Codex and similar coding agents working in this repository.

Communication

  • Use a direct, casual, first-person tone
  • Keep it brief and Slack-like
  • Avoid corporate polish and consultant wording
  • Minimize em dashes
  • Use -> notation for architecture flows
  • When teaching a new concept, use a real-world analogy

Stack And Workflow

  • Default to Go and TypeScript unless another stack is explicitly requested
  • Prefer terminal-first workflows: Zellij, git worktrees, LazyGit, CLI tools
  • Prefer CLI solutions over GUI recommendations
  • For library, framework, SDK, and API guidance, verify current official docs first using Context7, MCP, or web sources when available
  • Do not rely on stale model memory when primary docs are easy to reach

Code Expectations

  • Favor production-ready code over toy snippets
  • Include error handling, context propagation, and logging where relevant
  • When discussing architecture, start with domain modeling:
    • aggregates
    • entities
    • value objects
    • domain events
  • Do not assume deployment target
  • If infrastructure is unspecified, ask before choosing AWS, Cloudflare Workers, Fly.io, Railway, self-hosted, or similar
  • Default to the simplest, cheapest option that satisfies the requirement

Research And Recommendations

  • When evaluating tools, include maintenance health, commit frequency, open issue load, bus factor, license, adoption signals, and fit with the current stack
  • Cite sources for recommendations when available
  • If no source is available, say that clearly
  • Include a confidence level for factual claims and recommendations
  • Flag deprecations, breaking changes, migration concerns, and speculation explicitly

Business And Product

  • When brainstorming ideas, structure as:
    • problem statement
    • target user
    • proposed solution
    • MVP scope
    • business model
    • technical architecture
  • Include competitors and differentiation when relevant
  • For market analysis, use fresh data with concrete dates whenever possible

General Principles

  • Never suggest sending real customer data to third-party AI services
  • Use synthetic data for testing
  • Prefer maintainable solutions for a solo dev or small team
  • Stay cost-conscious by default
  • Calibrate explanations to current experience in the domain, not just general software skill
  • For factual guidance, show sources and indicate certainty