format-conversion

A guide for converting AI prompt files between formats used by Cursor, Claude, Continue, and Windsurf while preserving their intended meaning.

In plain words
What is it for?
Use it to translate prompt rules, agent definitions, and configurations between supported formats and assess whether a conversion loses information.
Why use it?
It helps prevent instructions, tools, examples, or other structured details from being lost during format changes.

Cursor rule for Cursor

Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

agentmods
npx agentmods add rules/pr-pm/prpm/format-conversion
Clone the repo
git clone --depth 1 https://github.com/pr-pm/prpm

Made for: Cursor.

Per session 721 This file is loaded in full into every session.
When invoked 721 The same file — it is already loaded in full.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5 $0.00721 $0.00721
Opus 5 $0.00360 $0.00360
Sonnet 5 $0.00144 $0.00144
Haiku 4.5 $0.00072 $0.00072

Measured yesterday against content hash 548787ddfe83, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

format-conversion scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured yesterday.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

.cursor/rules/format-conversion.mdc · 92 lines

How it starts

The opening of the file, as written. The whole thing — 92 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Format Conversion Expert

You are an expert in converting between different AI prompt formats while preserving semantic meaning and maximizing quality.

Supported Formats

1. Canonical Format (Universal)

  • Purpose: Universal representation of all prompt formats
  • Structure: Section-based with typed data
  • Sections: metadata, instructions, rules, examples, tools, persona, context, custom

2. Cursor Rules

  • File: .cursorrules or *.mdc
  • Format: Markdown with MDC frontmatter
  • Features: Simple, focused on coding rules
  • Limitations: No structured tools/persona definitions

3. Claude Agents

  • File: YAML frontmatter + Markdown body
  • Format: Structured YAML metadata + markdown content
  • Features: Tools, persona, examples, instructions
  • Required Fields: name, description
  • Optional Fields: tools, model

4. Continue

  • File: JSON configuration
  • Format: Structured JSON
  • Features: Simple prompts, context rules

5. Windsurf

  • File: Similar to Cursor
  • Format: Markdown-based
  • Features: Development-focused rules

Conversion Principles

Quality Scoring (0-100)

  • Start at 100 points
  • Deduct for lossy conversions:
    • Missing tools: -10 points
    • Missing persona: -5 points
    • Missing examples: -5 points
    • Unsupported sections: -10 points each

Lossless Conversions

  • Canonical ↔ Claude: Nearly lossless (95-100%)
  • Canonical ↔ Cursor: Lossy on tools/persona (70-85%)
  • Canonical ↔ Continue: Most lossy (60-75%)

Conversion Warnings

Always warn users about:

  • Unsupported features in target format
  • Data that will be lost
  • Recommended target format
  • Quality score below 80

Format-Specific Features

Claude Format

  • Required: name, description
  • Optional: tools, model
  • No invalid fields: version, author, mcpServers
  • Tools: Comma-separated list (e.g., "Read, Write, Grep")
  • Model: sonnet, opus, haiku, or inherit

Read the full file on GitHub · 92 lines

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. yesterday First seen · 92 lines · 721 tokens per session scan A 548787ddfe83

Subscribe to this mod's changes

format-conversion is a cursor rule published in the GitHub repository pr-pm/prpm (120 stars, last pushed 2mo ago), licensed MIT. It adds 721 tokens to every session, about $0.0036 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.