Data Export & Import

RedAmon provides multiple ways to export scan data, back up projects, and import previously exported projects. This page covers all export and import capabilities.

Scan Data Downloads

Each scanning module produces downloadable results from the Red Zone toolbar:

Download buttons only appear after the respective scan has completed.

Versioned graph snapshots. Scan Timeline keeps each full scan as an immutable graph version. The Recon Delta view can export a diff between any two versions as JSON (recon-delta-{from}-to-{to}.json), and the Version Manager lets you save the current graph as a version on demand.

Data Table Export

Every page-level table on the Attack Surface Graph supports three export formats: XLSX, JSON, and Markdown (MD). The three buttons sit next to each other in the table header bar.

All three buttons honour the current search/filter state — only visible/filtered rows are exported.

Tables that expose all three buttons

<ts> is YYYY-MM-DD-HH-MM-SS (UTC). For Node Inspector, <nodetype> is the lower-cased, dash-normalised type slug (e.g. domain, chain-step).

Node Inspector exports are scoped to the currently selected node type AND honor the user's column-visibility preferences — only visible columns are written, in display order (Name, then dynamic property columns alphabetically, then In/Out if visible). The JSON variant additionally includes a header object: { nodeType, generatedAt, columns, rows }. The Markdown variant uses # <NodeType> — Node Inspector Export as the title.

Step-by-step

1. Open the project on the Attack Surface Graph
2. Switch to the Table view using the view tabs below the toolbar
3. (Optional) pick a sub-tab — Node Inspector (default), All Nodes, JS Recon, or any Red Zone analytic
4. (Optional) filter the data using the search bar or node-type chips
5. Click XLSX, JSON, or MD in the header bar of the table
6. The browser downloads the file immediately — a toast confirms success or surfaces the error reason

Note: XLSX export validates the workbook before it leaves the browser; if a row contains data Excel cannot represent (e.g. corrupt UTF-8, raw binary), the file is repaired automatically rather than silently producing a file Excel refuses to open.

AI Agent Session Export

You can export any AI agent conversation as a Markdown report:

1. Open the AI Agent Drawer on the Red Zone
2. Load the conversation you want to export (or use the current one)
3. Click the download icon in the drawer header
4. A .md file downloads containing:
   • All user messages and agent responses
   • Thinking/reasoning steps
   • Tool executions with output
   • Findings and recommendations
   • Todo list states

Project Export

Export an entire project as a ZIP archive for backup or migration:

1. Go to the Projects page
2. Click on the project card to open its settings
3. Click the Export button
4. A ZIP file downloads containing:
   • Project configuration (all 180+ settings)
   • Conversation history with all messages
   • Neo4j graph data (all nodes and relationships)
   • Artifact files (scan results JSON files)
   • The Rules of Engagement document and its superseded versions (engagement/roe-revisions.json, the ten newest)

Plugin data and plugin skill versions are not part of a project export: they belong to the plugin installation, not to the project. The archive's manifest lists them as excluded, and an import never reads plugin data. To keep a plugin's data, use Export data in the plugin's details on the Plugins page.

Project Import

Import a previously exported project:

1. Go to the Projects page
2. Click the Import Project button
3. Select the ZIP file from a previous export
4. The project is recreated with all its data:
   • Configuration settings are restored
   • Conversations are recreated
   • Graph nodes and relationships are re-imported
   • Artifacts are restored
   • The Rules of Engagement document and its history are restored

Mutes travel with the graph. A muted finding stays muted after an import, with who muted it and why. A mute an external agent made over MCP also keeps the prefix of the access token that made it (muted_token), which is only the first characters of the token, never the token itself. In the imported project those prefixes belong to the exporting account's tokens, so the Token filter in Muted Nodes lists them even though no token of yours matches.

A Multi mute keeps its batch id (mm- followed by eight hex characters) in the same muted_token field. It is a random label that groups the findings muted from one suggestion, not a credential, so it is harmless in an export. In the imported project it still groups those mutes in Muted Nodes.

Jev engine choices are kept. A recon AI hook set to the TypeSafe Jev engine in the exported project stays on Jev after the import, even when your account has no Jev token. It then uses its static fallback at scan time, and the project form shows a No Jev token badge on it until you add a token or switch the hook back to the LLM. The token itself is never part of an export.

An imported RoE is checked like an upload. Each superseded document is held to the upload limits and its hashes are recomputed from its bytes. Every stored parse, the current document's and each superseded one's, is re-checked against this install's settings registry, so a bundle cannot plant a setting that a document may not set. The person who superseded each document belonged to the exporting install, so the import records you instead.

Note: On import, a new project ID is generated. The imported project will not conflict with existing projects even if the original still exists.

Neo4j Direct Access

For advanced users, the Neo4j graph database is directly accessible:

• Browser UI: http://localhost:7474 — Neo4j's built-in web interface for running Cypher queries
• Bolt protocol: bolt://localhost:7687 — for programmatic access from scripts or applications

Both ports are published on 127.0.0.1 only (not the LAN); reach them from the host.

Credentials: neo4j / your NEO4J_PASSWORD from .env. On a fresh install this is a strong value generated by redamon.sh; there is no insecure fallback, Compose refuses to start Neo4j if the variable is unset (${NEO4J_PASSWORD:?...}, STRIDE S13). Get it with grep '^NEO4J_PASSWORD=' .env.

Next Steps

• Red Zone — learn about the Data Table view
• Scan Timeline — version, diff, and schedule recon scans
• Troubleshooting — resolve common issues