> ## 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.

# Concepts Overview

> The canonical domain model that every Parmana artifact is built from

Parmana's domain model is a chain of immutable artifacts. Each artifact answers a specific question, and each one contributes to the next, from "who decided" through to "can this be independently proven."

```
Authority
    |
    v
Authorization
    |
    v
Intent
    |
    v
Policy Reference
    |
    v
Business Transaction
    |
    v
Decision
    |
    v
Execution
    |
    v
Evidence
    |
    v
Verification
    |
    v
Receipt
    |
    v
Execution Trust Record
```

## Reading the chain

| Artifact                                                        | Answers                                                |
| --------------------------------------------------------------- | ------------------------------------------------------ |
| [Authority](/docs/concepts/authority)                           | Who decided?                                           |
| [Authorization](/docs/concepts/authorization)                   | Was this request permitted to be evaluated?            |
| [Intent](/docs/concepts/intent)                                 | What should happen?                                    |
| [Policy](/docs/concepts/policy)                                 | What rules govern this action?                         |
| [Business Transaction](/docs/concepts/business-transaction)     | What was submitted for evaluation?                     |
| [Decision](/docs/concepts/decision)                             | Was the intent approved against policy?                |
| [Execution](/docs/concepts/execution)                           | What actually happened?                                |
| [Verification](/docs/concepts/verification)                     | Can the recorded execution be trusted?                 |
| [Receipt](/docs/concepts/receipt)                               | What's the cryptographic attestation of that trust?    |
| [Execution Trust Record](/docs/concepts/execution-trust-record) | What's the complete, immutable evidence package?       |
| [Replay](/docs/concepts/replay)                                 | Can this outcome be deterministically reproduced?      |
| [Override](/docs/concepts/override)                             | Was this record corrected after the fact, and by whom? |

## Two invariants worth internalizing early

**Decisions never execute.** Policy evaluation and execution are separate responsibilities, handled by separate components. A `Decision` records an outcome; it never performs an action.

**Nothing is ever mutated.** Every artifact in the chain is immutable once created. There is no "edit" operation anywhere in the domain model. Corrections are represented as new, additional artifacts (see [Override](/docs/concepts/override)) — the original record is never rewritten.