> ## Documentation Index
> Fetch the complete documentation index at: https://docs.manthan.systems/llms.txt
> Use this file to discover all available pages before exploring further.

# System Components

> Core components of the Parmana Execution Authorization architecture and how they interact

# System Components

Parmana is composed of modular components that work together to enforce deterministic authorization of execution.

Each component has a strict responsibility boundary.

***

## System Architecture

```text theme={null}
AI Systems → Signal Processing → Verified Signals → Governance → Authorization Decision → Execution Runtime → Attestation → Verification
1. AI Systems

AI systems generate raw outputs such as:

recommendations
predictions
proposed actions

AI systems do NOT participate in authorization decisions.

2. Signal Processing Layer

Transforms raw AI outputs into structured signals.

Responsibilities:

normalization
schema enforcement
validation preparation
3. Verified Signals

Validated signals that have passed integrity checks.

Only Verified Signals are allowed into Governance.

4. Governance (@parmanasystems/governance)

The deterministic policy evaluation engine.

Responsibilities:

policy evaluation
rule enforcement
authorization computation

Output: Authorization Decision

5. Authorization Decision

A deterministic result produced by Governance.

Possible outcomes:

APPROVE
REJECT
REQUIRE_OVERRIDE

This is the only input allowed for execution.

6. Execution Runtime (@parmanasystems/execution-runtime)

Enforces Authorization Decisions.

Responsibilities:

replay protection (executionId)
execution validation
deterministic enforcement
7. Attestation Layer

Cryptographically signs execution outcomes.

Includes:

decision hash
signal hash
policy version
runtime identity
8. Verification Layer

Allows external parties to independently verify:

Authorization Decisions
execution integrity
full system traceability

No system access is required.

Component Properties
Deterministic

All components produce reproducible outcomes.

Isolated responsibilities

Each layer has a strict functional boundary.

Verifiable

All outputs are cryptographically verifiable.

Secure by design

Invalid inputs are rejected at earliest stage.

Summary

Parmana is a layered system:

AI generates signals
Signal Processing structures them
Governance evaluates them
Execution Runtime enforces decisions
Attestation guarantees integrity
Verification ensures external trust
```