proompts

A curated set of Markdown prompts for AI-assisted product development, regulatory workflows, and general operations. Prompts are organized by topic—ranging from code reviews to market research—so you can mix and match them in your own agentic workflows.

Docs

• docs/ – additional docs and a full table of contents

Prompt Schema

Prompts are stored as .json files that conform to docs/prompt_schema.json. Each file contains these top-level fields:

• id – unique kebab-case identifier
• title – short descriptive title
• category – directory or topical grouping
• author – optional author attribution
• created – date added (YYYY-MM-DD)
• last_modified – most recent modification date
• tested_model – optional model/version used when testing
• temperature – optional sampling temperature
• tags – optional array of keywords
• prompt – object holding the prompt text

The prompt object must include purpose, context, instructions, inputs, and output_format. Optional keys are additional_notes, example_usage, and references. See docs/template_prompt.json for a filled-out example.

The body of the prompt mirrors the template in prompt_tools/L5_standardize-prompt-files.md with sections for Purpose, Context, Instructions, Inputs, Output Format, Additional Notes, Example Usage, and References.

Validation

Run the helper script to verify JSON formatting and the docs index before committing:

./scripts/validate_json.sh

The script requires jq and Python 3. It checks that docs/index.md and docs/table-of-contents.md are current and then runs jq against every *.json file. The same check runs in GitHub Actions. This repository also runs workflows to generate missing overview.md files, verify file naming, and commit the docs index when it changes.

Contributing

1. Create prompts as .json files that follow docs/prompt_schema.json and place them in the appropriate folder.
2. Before committing, sanitize and standardize the file using prompt_tools/L5_prompt_sanitiser.md and prompt_tools/L5_standardize-prompt-files.md.
3. Run scripts/validate_json.sh to verify JSON formatting and update the docs index.
4. Optionally, run prompt_tools/01_architecture_review_pipeline.md to audit the repository.
5. If you create a new directory, an overview.md will be generated automatically by the workflow.
6. The same validation runs in CI, but running scripts/validate_json.sh locally helps catch issues early.

License

This project is licensed under the Proompts Personal Use License. Individuals may freely use, modify, and distribute the prompts for personal, non-commercial purposes. Commercial entities must obtain written permission from Frederick de Ruiter before using the material.

This site is open source. Improve this page.