type: prompt-template name: prompt-system-guide model: anthropic/claude-haiku-4.5

Explain how QNTX prompt templates work, using {{subject}} as the example.

Include:

  1. Frontmatter: YAML configuration at the top (type, name, model)
  2. Template syntax: {{field}} placeholders map to attestation fields
  3. X-sampling: Test prompts against X random attestations before deployment
  4. Available fields: {{subject}}, {{context}}, {{attributes}}, {{predicates}}, etc.
  5. Preview workflow: Edit template → Run Preview → Compare outputs → Deploy

Show how {{subject}} demonstrates these concepts, referencing its implementation in {{context}}.


QNTX Prompt System Development Tracker

Vision: Enable confidence in prompt changes through X-sampling preview: sample X attestations, execute prompt X times, compare outputs against previous versions. This transforms prompt development from risky production experiments to controlled, iterative refinement.

Tracking PR: #301 - QNTX Prompt: Attestation-based LLM prompt templates with frontmatter

✅ Merged Work

🔥 Critical Issues (Next PRs)

📋 Non-Critical Issues (Future)


Developer Workflow

  1. Pick an issue from the lists above
  2. Implement the feature in a branch
  3. Update this README when done:
  4. Execute this README as a prompt to verify the documentation explains the feature

This README is itself an executable prompt - it should always accurately explain and demonstrate QNTX prompt capabilities.