kata-pov-system-prompt

To generate a pre-operational system prompt, the workflow reads your agent's pov.md file and outputs a system-prompt.md containing Identity, Capabilities, Restrictions, and Style blocks with an explicit stage declaration.

A pre-operational PoV system prompt is a structured prompt defining an agent's persona scope before deployment. You need it when translating persona vision, value metrics, and out-of-scope restrictions into a constrained prompt for agent construction.

Structure system prompt capabilities and restrictions by defining a maximum of three scope-constrained capabilities aligned to your primary use case, followed by explicit out-of-scope restrictions to ensure safety and consistency.

Yes, system prompt generation includes prompt injection validation by invoking an adversarial validation step in minimum-viable mode before persisting the final system-prompt.md file.

No, you cannot use placeholders in your agent system prompt file. The generation workflow ensures all placeholders are removed and the literal stage: pre-operational text is included before the prompt is persisted.

Limitations of minimum viable system prompts include strict capability limits of three functions, mandatory pre-operational stage declarations, and required adversarial validation before persistence to ensure safe agent behavior.