Metadata-Version: 2.4
Name: rpek-lite
Version: 1.0.1
Summary: RPEK Prompt Execution Kernel - Certified Baseline
Author: RPEK Governance Board
Classifier: Development Status :: 5 - Production/Stable
Classifier: Intended Audience :: Developers
Classifier: Programming Language :: Python :: 3
Classifier: Programming Language :: Python :: 3.8
Classifier: Programming Language :: Python :: 3.9
Classifier: Programming Language :: Python :: 3.10
Classifier: Programming Language :: Python :: 3.11
Classifier: Programming Language :: Python :: 3.12
Classifier: License :: OSI Approved :: MIT License
Classifier: Operating System :: OS Independent
Classifier: Topic :: Software Development :: Build Tools
Requires-Python: >=3.8
Description-Content-Type: text/markdown
License-File: LICENSE
Requires-Dist: pytest==7.4.3
Dynamic: author
Dynamic: classifier
Dynamic: description
Dynamic: description-content-type
Dynamic: license-file
Dynamic: requires-dist
Dynamic: requires-python
Dynamic: summary

# Prompt Governance RPEK-Lite

> [!WARNING]
> RPEK-LITE-V2 enforces live evidence for all TASK_COMPLETION reports. Simulated timestamps, log-inferred results, and heartbeat confirmations are auto-rejected.

A standalone, reusable, lightweight prompt execution kernel. This package scaffolds a generic framework for governing, building, and executing AI prompt tasks in a predictable, auditable manner.

**Core Principle:** No raw prompts. All prompts must pass through RPEK-LITE governance: adapter context, task payload, and strict validation.

## Installation

RPEK Lite is distributed as a standard Python package.

```bash
# Install directly from PyPI
pip install rpek-lite
```

This installs the package and registers the `rpek-build`, `rpek-task`, and `rpek-admit` CLI commands globally.

## Quick Start (Governed Workflow)

RPEK Lite enforces a strict pipeline for prompt execution: **Task Definition -> Prompt Build -> Admission -> Execution Evidence**.

1. **Initialize Project:** Create your consumer directory (e.g., `projects/my_project`) and an `adapter.json`.
2. **Scaffold a Task (`rpek-task`):**
   ```bash
   rpek-task my_project TASK_001 \
     --objective "Fix the login bug" \
     --allowed "auth/login.py"
   ```
3. **Build the Prompt (`rpek-build`):**
   ```bash
   rpek-build my_project TASK_001
   ```
   This compiles `build/final_prompt_TASK_001.md`.
4. **Admit Intent (`rpek-admit`):**
   Before executing the payload against an LLM, confirm admission:
   ```bash
   rpek-admit --project my_project --task-id TASK_001 --intent-text "Implement login bugfix"
   ```
5. **Execute & Record Evidence:** Submit the compiled prompt payload to your language model, and retain the RPEK-Lite generated execution JSON in your `audit/` directory as canonical evidence.

## CLI Commands Reference

- `rpek-task`: Scaffolds a new governed task definition (`task.json`).
- `rpek-build`: Compiles the task and adapter into a final Markdown prompt payload, leaving an audit trace.
- `rpek-admit`: Runs the canonical ingress controller admission decision for the specified task intent.
