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

# Security

> Protecting your sensitive information

Chatledger is designed with security in mind. Since AI conversations often involve pasting API keys, credentials, or proprietary code, the extension includes robust features to prevent accidental leaks.

## Local Processing

<Check>
  All processing happens locally on your machine. Chatledger communicates only
  with the local Antigravity process (localhost). No data is sent to external
  servers.
</Check>

## Sensitive Content Detection

The extension scans every conversation before export using a set of regular expressions.

### Built-in Patterns

Chatledger detects:

* OpenAI API Keys (`sk-...`)
* AWS Access Keys & Secrets
* Private Keys (PEM/OpenSSH)
* Generic "password=" or "secret=" assignments
* JWT Tokens
* Bearer Tokens

### Custom Patterns

You can add your own patterns in `settings.json`:

```json theme={null}
"chatledger.additionalSensitivePatterns": [
  "sk-prod-[a-zA-Z0-9]{20}",
  "INTERNAL_SECRET_\\w+"
]
```

## Activity Bar Security View

The **Chatledger Security** view (shield icon in the Activity Bar) provides a centralized dashboard for all detected issues.

* **Review**: See a list of files containing potential secrets.
* **Navigate**: Click on a match to jump effectively to the line in the exported file.
* **Dismiss**: If a match is a false positive, you can dismiss it.

## Censoring

If you enable `chatledger.censorSensitiveContent`, the extension will automatically redact detected secrets in the exported Markdown files.

* **Original**: `const apiKey = "sk-12345abcdef";`
* **Censored**: `const apiKey = "***CENSORED***";`

<Warning>
  Censoring is irreversible in the exported file. The original conversation in
  Antigravity remains untouched.
</Warning>