Building scalable AI agents with modular prompt transpilation

Summary

Monolithic system prompts become difficult to reason about, test, and change as teams add safety policies, domain rules, formatting requirements, and escalation behaviors. The proposed approach treats prompts as software artifacts: modular skill files are composed with includes, variables, and macros, then transpiled into deterministic artifacts. Build-time validation checks missing imports, undefined variables, and circular dependencies, while dependency graphs and CI/CD golden-file comparisons catch drift before deployment. At runtime, progressive disclosure keeps a stable control plane loaded while agents retrieve task-specific skills, reducing context exhaustion and noise. Agents can propose new skill modules or import changes through pull requests, but human review and evaluation remain part of the merge process.

Takeaways

1. 1

   A prompt fragment can be modeled as a node in a dependency graph, allowing recursive imports and other dependency errors to be detected during the build.

2. 2

   CI can regenerate a transpiled prompt and compare it with the committed golden file; differing outputs cause the build to fail and expose source-artifact drift.

3. 3

   Agents may draft new skill modules and import changes as pull requests, with the transpiler, evaluations, and human review applied before a change is merged.