🏰 Palisade

Enterprise-grade ML model security scanner. Detects backdoors, supply chain attacks, and malicious payloads before they hit production.

Powered by a high-performance Rust core, Palisade delivers maximum speed and memory efficiency, enabling it to scan 70B+ parameter models on standard hardware.

⭐ Key Capabilities

• Blocks Pickle RCE - Completely prevents remote code execution via pickle files.
• Validates Model Integrity - Verifies SafeTensors and GGUF formats against tampering.
• Verifies Supply Chain - Enforces Sigstore signatures, SLSA provenance, and generates ML-BOMs.
• Catches Injection Attacks - Prevents tokenizer hijacking, config manipulation, and metadata exploits.
• Zero-Trust Architecture - Treats all models as potentially malicious until verified.

15 Security Validators provide multi-layered defense in depth (10 universal + 5 format-specific).

Commands

Command	Description
scan	Scan model file(s) or directory for security threats (static analysis).
verify-sigstore	Verify Sigstore model transparency signature.
verify-slsa	Verify SLSA provenance attestation.
track-provenance	Generate comprehensive provenance tracking report.
policy	Policy template management.

Using Palisade in GitHub Action

The Palisade GitHub Action enables you to run Palisade as part of your CI/CD workflows. It installs Palisade from PyPI inside a Docker container and runs the selected Palisade command on your repository, models, or artifacts.

name: Palisade Security Scan

on:
  pull_request:
  push:
    branches: [ "main" ]

jobs:
  palisade:
    name: Run Palisade Security Scan
    runs-on: ubuntu-latest

    steps:
      - uses: actions/checkout@v4

      - name: Run Palisade scan
        uses: highflame-ai/palisade-githubaction@v1
        with:
          args: "scan . --format sarif --output palisade.sarif"

      - name: Upload SARIF to GitHub Code Scanning
        uses: github/codeql-action/upload-sarif@v3
        with:
          sarif_file: palisade.sarif

Any Palisade command can be executed by passing the appropriate value to the args input.

Performance

Palisade uses a native Rust core to handle massive models efficiently without OOM errors. It employs smart streaming and memory mapping to validate models larger than available RAM.

Model Size	Format	Scan Time	Memory Usage	Validators
511.38 MB (250M)	SafeTensors	3.7 s	115.4 MB	13
2.09 GB	SafeTensors	14.3 s	115.4 MB	13
3.8 GB (7B Q4_K_M)	GGUF	29.4 s	140 MB	11
9.4 GB	SafeTensors	74.3 s	119.4 MB	13

Palisade Ecosystem

• 🏆 Palisade Leaderboard
  Explore real-world scan performance, validator coverage, and benchmark results.
  👉 SafeLLM-leaderboard

• 📘 Palisade Documentation
  Learn about validators, supported formats, CLI usage, and security guarantees.
  👉 Highflame docs

🏰 Built with ❤️ by highflame • Securing the LLM supply chain