CIS Benchmark Auditor

Automated security auditing tool for CIS Benchmarks. Runs checks against a local machine or a remote host via SSH and produces a colour-coded XLSX report showing passed, failed, and manual-review items with remediation guidance.

Supported benchmarks

Benchmark	Version	Checks	Implementation
CIS NGINX	v3.0.0	44	Hand-coded Python functions
CIS Debian Linux 13	v1.0.0	343	Data-driven XLSX parser

Requirements

pip install -r requirements.txt

paramiko is only needed for SSH mode. The rest of the standard library covers local execution.

Usage

List available profiles

python main.py nginx --list-profiles
python main.py debian --list-profiles --benchmark-file CIS_Debian_Linux_13_Benchmark_v1.0.0.xlsx

Audit local machine

python main.py nginx  --profile "Level 1 - Webserver"

python main.py debian --profile "Level 1 - Server" \
    --benchmark-file CIS_Debian_Linux_13_Benchmark_v1.0.0.xlsx

Audit remote host via SSH

python main.py debian \
    --mode ssh --host 10.0.0.5 --user admin --key ~/.ssh/id_rsa \
    --profile "Level 1 - Server" \
    --benchmark-file CIS_Debian_Linux_13_Benchmark_v1.0.0.xlsx

Run specific checks only

python main.py debian \
    --profile "Level 1 - Server" \
    --benchmark-file CIS_Debian_Linux_13_Benchmark_v1.0.0.xlsx \
    --checks 1.1.2.1.1,2.1.1,5.1.1

All options

positional:
  benchmark             nginx | debian

options:
  --benchmark-file XLSX Path to benchmark XLSX (required for: debian)
  --mode {local,ssh}    Execution mode (default: local)
  --host HOST           SSH host
  --port PORT           SSH port (default: 22)
  --user USER           SSH username (default: root)
  --key KEY             Path to SSH private key
  --password PASSWORD   SSH password
  --profile PROFILE     Profile to audit
  --checks CHECKS       Comma-separated check IDs
  --list-profiles       Print profiles and exit
  --output XLSX         Output report path (default: auto-named)
  --verbose, -v         Print check progress

Report format

The XLSX output contains four sheets:

Summary — total counts, timestamp, target host
Passed — checks that passed with actual vs expected values
Failed — checks that failed with actual vs expected values and remediation steps
Manual Review — checks that require human judgment, with collected command output
Errors (if any) — checks that could not be executed

Project structure

main.py                         # CLI entry point (all benchmarks)
requirements.txt
cis_auditor/
  core/
    models.py                   # CheckResult, AuditReport dataclasses
    executor.py                 # LocalExecutor, SSHExecutor
    report.py                   # XLSX report generator
  benchmarks/
    base.py                     # CheckDefinition, BenchmarkInfo
    nginx/
      __init__.py               # Wraps hand-coded nginx_audit checks
    debian/
      __init__.py               # Lazy-load wrapper
      loader.py                 # XLSX parser → CheckDefinition objects
nginx_audit/                    # Hand-coded NGINX check functions (44 checks)
  checks.py
  executor.py
  models.py
  report.py
  runner.py

Adding a new benchmark

Create cis_auditor/benchmarks/<name>/
Implement loader.py (data-driven from XLSX) or checks.py (hand-coded)
Expose load(xlsx_path) -> BenchmarkInfo in __init__.py
Add the name to BENCHMARK_NAMES in main.py

The CheckDefinition.run() method handles all four pass-condition types (empty_output, non_empty_output, contains_pass, manual) automatically, so new benchmarks only need to supply the audit script and the right condition.

3.8 KiB Raw Blame History