Prompts
# Phase 5 - AI Prompt Orchestration
> *"Generic prompts produce generic output. Precise prompts produce production-grade output."*
---
## Activation
This phase activates when the documentation baseline is established.
The orchestrator's job here is not to execute tasks -
it is to generate the exact prompts that executing agents will receive.
Every prompt must be:
- Role-specific (who is this agent acting as?)
- Context-injected (what does it need to know?)
- Constraint-defined (what must it never do?)
- Output-specified (what exactly should it produce?)
- Self-validating (how does it verify its own output?)
A prompt without injected context is a prompt that will hallucinate.
---
## Prompt Construction Template
Every generated prompt follows this structure:when to use it
Community prompt sourced from the open-source GitHub repo a7medalyapany/kavro (MIT). A "Prompts" style prompt — adapt the placeholders and specifics to your task. Imported as-is and not independently retested here, so check the output before relying on it.
tags
careercommunitygeneral
source
a7medalyapany/kavro · MIT