AI Agents Make Best Practices Non‑Optional

Read Articleadded Dec 29, 2025
AI Agents Make Best Practices Non‑Optional

AI coding agents work best when ambiguity is eliminated and guardrails are strict. The author’s team enforces 100% code coverage, clear file organization, fast ephemeral concurrent environments, and end‑to‑end typing with automated linters and generated clients. These investments turn best practices from “nice to have” into essential infrastructure that lets agents reliably produce correct code.

Key Points

  • Require 100% code coverage to force agents to validate the behavior of every line they change, creating a clear to‑do list and reducing ambiguity.
  • Design the filesystem as an interface: use clear namespaces and many small, focused files to improve agent navigation and context loading.
  • Make dev environments fast, ephemeral, and concurrent so agents can iterate in tight loops, spin up clean contexts instantly, and run in parallel without conflicts.
  • Automate strict quality gates: linters/formatters with auto‑fixes and typed languages (the author favors TypeScript) to reduce the model’s search space and encode intent.
  • Adopt end‑to‑end typing: OpenAPI‑generated clients, Postgres types/checks/triggers, and typed query builders (e.g., Kysely), wrapping third‑party clients for strong types.

Sentiment

Overall, the sentiment of the Hacker News discussion is mixed to moderately skeptical. While some commenters agree with the core premise that AI changes the value proposition of certain development practices and can benefit from strict environments, a significant portion expresses strong reservations and skepticism regarding the effectiveness of metrics like 100% coverage, the general capabilities of AI for 'good code,' and the implications of the article's recommendations.

In Agreement

  • 100% code coverage, while not guaranteeing zero bugs, is valuable for AI agents because it maximizes their ability to 'crank out code' by providing clear, easily verifiable targets and ensuring every line is exercised.
  • Tests are crucial for AI agents, acting as 'reified context' that makes future agentic interactions safer, even if the AI doesn't need them as much as humans for initial scaffolding.
  • Restricting degrees of freedom through strict practices like strong typing and structured environments is absolutely critical for being productive and safe with AI agents at scale.

Opposed

  • 100% code coverage is a misleading metric; it can be easily gamed to cover code that still breaks in most paths, failing to guarantee correctness or robustness (e.g., a function tested only for a single successful input).
  • The premise that AI 'forces' good code might be a cynical take or 'drinking the Kool-Aid,' implying AI introduces problems that then necessitate these safeguards, akin to praising drunk-driving for increased seatbelt use.
  • Inexperienced programmers might misinterpret the article's ideas as universally sound best practices, overlooking nuances or potential downsides.
  • LLMs' ability to stay on top of genuinely new design concepts, languages, or truly innovate is questionable, and using them to write 'good code' might pose security risks; they might be better suited for prototypes or 'bad code'.
  • There's a general distrust of advice from bloggers, particularly those with polished sites or commercial interests, suggesting their recommendations might lack objective value.
AI Agents Make Best Practices Non‑Optional