> For the complete documentation index, see [llms.txt](https://docs.lpp-minduniverse.org/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.lpp-minduniverse.org/lingua-pactum-protocol-lpp-documentation/7.-mind-universe-governance-stack/7.2-sourcemind.md).

# 7.2 SourceMind

**SourceMind** is the source provenance and input trust governance component of Mind Universe.

Its current canonical public role is:

> **Determine where relevant information originated and what evidentiary or trust status should accompany it downstream.**

This places SourceMind at the beginning of the governance chain.

```
Source
        ↓
SourceMind
        ↓
Governed Input
        ↓
Semantic & Task Processing
```

#### Source Provenance

AI systems increasingly rely on information from:

* retrieval systems,
* enterprise databases,
* external documents,
* APIs,
* user uploads,
* data pipelines,
* training or fine-tuning datasets,
* and other agents.

A downstream model may be able to use the information without knowing whether the source was:

* authentic,
* authorized,
* manipulated,
* disputed,
* incomplete,
* or unknown.

SourceMind addresses this provenance problem.

The source layer may preserve information such as:

* source identity,
* source reference,
* origin metadata,
* timestamp,
* provenance fingerprint,
* authorization status,
* and other trust-relevant evidence.

#### Evidence Origin

The purpose is not only to determine whether information exists.

It is to preserve:

> **Where did this evidence come from?**

That distinction becomes important later when an Action Candidate depends on external evidence.

If the system cannot reconstruct the origin of a critical claim, downstream governance should not treat that claim as equivalent to verified evidence.

#### Input Trust

SourceMind may assign or preserve information about the trust status of input.

This does not mean that SourceMind decides whether an action is admissible.

Rather:

```
SourceMind
        ↓
produces source / trust evidence
        ↓
downstream governance may consume it
```

#### Earlier SourceMind Design Lineage

The dedicated SourceMind white paper developed a broader data-input governance model around a **Data Accountability Chain**.

That design included concepts such as:

* Source Verification Layer,
* Label Trust Engine,
* High-Impact Sample Protector,
* Bias & Manipulation Detector,
* and Data Accountability Chain API.

It also proposed source fingerprints using combinations such as:

```
Hash
+
Timestamp
+
Source ID
```

and lifecycle-oriented responsibility records.

These remain relevant design lineage for SourceMind.

However, GitBook v2.0 does not make every earlier SourceMind data-governance mechanism a mandatory LPP requirement.

The canonical architectural role is narrower:

> **Source provenance, evidence origin, and input trust.**

#### SourceMind Does Not Create Authority

A trusted information source is not the same as legitimate action authority.

Therefore:

```
Trusted Source
⇏
Authority Object
```

and:

```
Verified Evidence
⇏
ADMIT
```

SourceMind can strengthen the evidentiary foundation of a task.

It cannot authorize the task.

***