inalign

<p align="center">
  <h1 align="center">InALign</h1>
  <p align="center"><strong>Tamper-proof audit trails for AI coding agents</strong></p>
  <p align="center">Know what your AI agents did. Prove it. Cryptographically.</p>
</p>

<p align="center">
  <a href="https://pypi.org/project/inalign-mcp/"><img src="https://img.shields.io/pypi/v/inalign-mcp?color=blue" alt="PyPI"></a>
  <a href="https://github.com/Intellirim/inalign/blob/main/LICENSE"><img src="https://img.shields.io/github/license/Intellirim/inalign" alt="License"></a>
  <a href="https://pypi.org/project/inalign-mcp/"><img src="https://img.shields.io/pypi/pyversions/inalign-mcp" alt="Python"></a>
  <a href="https://modelcontextprotocol.io"><img src="https://img.shields.io/badge/MCP-compatible-green" alt="MCP"></a>
</p>

---

## The Problem

AI coding agents (Claude Code, Cursor, Cursor) can read, write, and execute anything on your machine. When something goes wrong, you have no reliable way to answer:

- **What** did the agent actually do?
- **Who** told it to do that?
- **When** did it happen?
- **Can I prove it** to my team, auditors, or a court?

Logs can be edited. Memory fades. You need a chain of evidence that **cannot be tampered with**.

## The Solution

InALign is an open-source [MCP server](https://modelcontextprotocol.io/) that sits inside your AI agent and records every action into a **SHA-256 hash chain** -- each record cryptographically linked to the previous one. Modify any record and the chain breaks. Immediately detectable.

```
User prompt ──> Agent action ──> InALign records it
                                  |
                                  +-- SHA-256 hash chain (tamper-proof)
                                  +-- W3C PROV knowledge graph (ontology)
                                  +-- Full conversation capture (prompts + responses)
                                  +-- 6-tab interactive HTML dashboard
                                  +-- GraphRAG risk analysis (11 MITRE ATT&CK