HomeDocsBlogDownload

ClaudeArmor v2.4.1: Forensic AI Threat Reports & Custom Threat Policies

We're thrilled to announce ClaudeArmor v2.4.1, our biggest release this quarter. This update brings three major features that our community has been requesting: Forensic AI threat explanations that tell you why something is dangerous, support for custom organization threat policies, and a significant performance boost across all scan engines.

What's New in v2.4.1

๐Ÿง 

Forensic AI Threat Reports

Every detected threat now includes an executive explanation generated by ClaudeArmor's proprietary Forensic AI Engine. Instead of raw heuristic IDs, you'll see exactly what the threat does, why it's dangerous, and specific remediation solutions for your codebase.

๐Ÿ“

Custom Enterprise Threat Policies

Define your own organization-specific policies alongside ClaudeArmor's built-in neural heuristics. Block unauthorized internal API endpoints, proprietary code exfiltration, or compliance-specific violations.

โšก

42% Faster Scans

We rewrote the file traversal engine using a parallel execution pool, enabling high-speed scanning of large repositories. A 10,000-file monorepo that previously took 8.4 seconds now completes in under 5 seconds.

Forensic AI Threat Reports in Action

When ClaudeArmor detects a threat, each finding includes a contextual forensic breakdown and step-by-step solution:

$ claudearmor scan ./my-project

๐Ÿ” Scanning 1,247 files...

โš ๏ธ  THREAT DETECTED: config/database.js (line 14)
   Policy: env_exfiltration_vector
   Severity: HIGH

   ๐Ÿ“‹ Forensic AI Analysis:
   This code reads DATABASE_URL from environment variables
   and passes it to an unvalidated HTTP request on line 18.
   An attacker who controls the endpoint could capture your
   database credentials.

   โœ… Remediation Solution:
   1. Validate the endpoint URL against an allowlist
   2. Use a secrets manager instead of env variables
   3. Enable ClaudeArmor Strict Workspace Boundary

Authoring Organization Policies

Create a .claudearmor/policies/ directory in your project root to enforce custom security boundaries across all developer workstations.

Custom rules integrate seamlessly with the AI threat report system โ€” your custom detections get the same plain-English explanations as built-in rules.

Performance Improvements

The scan engine rewrite delivers measurable improvements across every project size:

These benchmarks were measured on an M2 MacBook Pro with 16GB RAM. Performance gains are even more pronounced on machines with 8+ CPU cores.

How to Upgrade

# Update via npm
npm update -g claudearmor

# Or download the latest desktop app
# Visit https://claudearmorsec.com/download

# Verify your version
claudearmor --version
# โ†’ ClaudeArmor v2.4.1

For the full list of changes, see the changelog. If you encounter any issues, please open an issue on GitHub.

Activate ClaudeArmor v2.4.1 Today

Forensic AI threat reports, enterprise policies, and 42% faster scans.

View Defense Passes โ†’
๐Ÿ‘จโ€๐Ÿ’ป

Marcus Rivera

Engineering Lead at ClaudeArmor. Building developer tools that make security effortless. Previously at Datadog and Vercel.