PromptScript vs Manual Configuration

The Manual Approach

Without PromptScript, teams maintain separate configuration files for each AI tool in every repository. Even with just 2-3 tools, this quickly becomes unmanageable across many repos:

Files to maintain (per repo)Problems

.github/copilot-instructions.md   # GitHub Copilot
CLAUDE.md                          # Claude Code
.cursor/rules/project.mdc         # Cursor
AGENTS.md                          # Codex / Factory AI
GEMINI.md                          # Gemini CLI
... one file per tool

Problem	Impact
No single source of truth	Instructions drift between tools
No validation	Errors go undetected until runtime
No inheritance	Copy-paste across repos
No audit trail	Who changed what and when?
Manual updates	One policy change = PRs in every repo, for every tool
Vendor lock-in	Switching tools means rewriting instructions, READMEs, and docs
:::
::::

The PromptScript Approach

With PromptScript, you write one .prs file and compile to all 50 coding agent targets:

Single source file

@meta { id: "my-project" syntax: "1.5.0" }

@inherit @company/backend-standards

@identity {
  """
  You are an expert developer working on the API service.
  """
}

@standards {
  code: { languages: ["TypeScript"], testing: ["Vitest"] }
}

@restrictions {
  - "Never expose API keys"
  - "Always validate input"
}

One commandBenefits

prs compile

Generates all output files automatically:

.github/copilot-instructions.md
CLAUDE.md
.cursor/rules/project.mdc
.windsurf/rules/project.md
.clinerules
AGENTS.md
GEMINI.md
... all configured targets

Benefit	How
Single source of truth	One `.prs` file, up to 50 coding agent outputs
Compile-time validation	Errors caught before deployment
Hierarchical inheritance	Org → Team → Project
Full audit trail	Git history on `.prs` files
Controlled updates	Change registry, then run `prs update` and recompile in each repo
Tool-agnostic	Switch tools without rewriting
:::
::::

Side-by-Side Comparison

Aspect	Manual	PromptScript
Files per repo	5-50	2 source files plus generated outputs
Update a policy	100+ manual edits	Registry update, dependency update, recompile
Add new tool	Write new file format	`prs compile`
Validation	None	Compile-time + CI/CD
Inheritance	Copy-paste	`@inherit` / `@use`
Consistency	Hope for the best	Validated in CI
Onboarding time	Hours per tool	Minutes

Getting Started

Ready to switch? See the Migration Guide to convert existing configs, or start fresh with the Getting Started guide.

Evaluating for a team or organization? Also see:

Target Platform Matrix - exact output paths per tool
CI/CD Integration - validation and drift detection in your pipeline
Enterprise Setup - registry, policies, and governance across repositories