v6.0.0 — Goal Hierarchy, Knowledge Base, Meta-Learning, Markdown Skills

What's New in v6.0.0

Goal Hierarchy

Mission-driven planning: the agent pursues strategic objectives, not just "find vulns."

argus agent example.com --mission data_exfiltration
argus agent example.com --mission admin_access
argus agent example.com --mission rce

Knowledge Base

6 built-in exploit patterns the agent applies based on detected tech stack:

• WordPress CSRF chains
• GraphQL introspection → IDOR
• Laravel debug → .env leak
• JWT algorithm confusion
• SSRF via redirect
• File upload bypass

Meta-Learning

Agent tracks skill effectiveness per technology and auto-adjusts strategy.

Markdown Skill System

Define custom skills as .md files that Argus auto-loads:

---
name: check_wordpress
description: WordPress-specific security checks
tools: [nuclei, httpx]
---
1. Probe /wp-admin for accessibility
2. Run nuclei with wordpress tags

Bug Fixes

• Finding severity: expanded from INFO|LOW to full 5-level scale
• Nuclei severity mapping was inverting HIGH to LOW
• DNSX targets were computed but never passed
• XSS severity: LOW → MEDIUM; SQLi: LOW → HIGH
• 11 critical bugs fixed, 63 new audit tests

Repository

• Professional README with banner and collapsible sections
• Full documentation suite (docs/)
• CONTRIBUTING, SECURITY, CODE_OF_CONDUCT, CHANGELOG
• GitHub issue/PR templates, CI workflow, Dependabot

Stats

• 808 tests passing
• 15 tools + 7 OSINT APIs
• 11 built-in skills + custom markdown skills

See CHANGELOG.md for full version history.

Argus v1.2.0 — Bulk Scanner (Mass Scanning Pipeline)

What's New in v1.2.0

🔍 Bulk Scanner — Three-Stage Mass Scanning Pipeline

Scan up to 50 hosts in a single session with automatic parallelization.

Stage 1 — Target Discovery

argus bulk targets.txt          # from file (one host per line)
argus bulk 192.168.1.0/24       # from CIDR range
argus bulk AS12345              # from ASN (uses bgpview.io, no key needed)
argus bulk --shodan "org:Acme"  # from Shodan query (requires ARGUS_SHODAN_KEY)
argus bulk targets.txt 10.0.1.0/24  # multiple sources combined

Stage 2 — Recon + Technology Identification

The new bulk preset runs per-target: dns → httpx → techstack → headers

Stage 3 — Automated Vulnerability Checking

nuclei + CVE correlation (NVD API) on every live host

Results

• Individual reports per target: ~/.argus-lite/scans/bulk-{id}/{host}/report.html
• Combined summary: ~/.argus-lite/scans/bulk-{id}/summary.html
  • Stats: targets / live hosts / findings / CVEs / highest risk
  • Per-host table with risk badges + links to individual reports
  • Technology distribution heatmap
  • CVEs found on multiple hosts

# Full example:
argus bulk targets.txt --preset web --concurrency 5 --output html

Technical Details

• asyncio.Semaphore(concurrency) — parallel scanning with configurable limit
• Partial failures isolated — one host failing doesn't abort the run
• TargetExpander: file / CIDR (ipaddress stdlib) / ASN (bgpview.io API) / Shodan
• BulkConfig: max_concurrent=5, max_targets=50 (configurable)

Tests

491 tests (40 new) — all passing

Argus v1.1.0 — CVE Correlation, TUI, Scan Templates, GitHub Actions, Docker

What's New in v1.1.0

🔍 CVE Correlation (NVD API)

Automatically queries the National Vulnerability Database after detecting technologies. Shows CVE IDs and CVSS scores in reports.

• Detects CVEs for any technology with a known version (WordPress 6.3.1, Apache 2.4, PHP 8.0, etc.)
• CVSS v3.1 / v3.0 / v2 score parsing
• Configurable via ARGUS_NVD_KEY (free API key increases rate limit from 5 to 50 req/30s)

🖥️ Interactive TUI

Beautiful real-time scan dashboard powered by Textual.

argus scan example.com --tui

• Left panel: live stage progress log
• Right panel: findings as they're discovered
• Press Q to exit after scan completes

📄 YAML Scan Templates

Declarative scan configuration — share and version-control your scan configs.

argus run examples/quick_scan.yaml --target example.com
# or: TARGET=example.com argus run examples/quick_scan.yaml

🤖 GitHub Actions Integration

Use Argus as a CI/CD security gate:

- uses: cortexc0de/argus-lite@v1
  with:
    target: ${{ vars.SCAN_TARGET }}
    preset: quick
    fail-on: HIGH
    output-format: sarif

Uploads SARIF reports to GitHub Security tab automatically.

🐳 Docker Image

Run on any platform without Kali Linux:

docker run -v ./reports:/reports ghcr.io/cortexc0de/argus-lite scan example.com --no-confirm

Test Coverage

• 451 tests (40 new) — all passing
• CVE enricher: 12 tests
• TUI: 8 tests
• Scan templates: 10 tests
• Exit code: 8 tests

Installation

# Kali Linux (native — recommended):
git clone https://github.com/cortexc0de/argus-lite
cd argus-lite && sudo ./install.sh

# Docker (any platform):
docker pull ghcr.io/cortexc0de/argus-lite:v1.1.0

Argus v1.0.0 — AI-Powered Security Scanner

Argus v1.0.0

AI-powered security scanner for Kali Linux with smart pipeline technology.

Highlights

• 14 integrated tools — subfinder, naabu, nuclei, httpx, katana, dnsx, tlsx, whatweb, ffuf, gau, gowitness + Shodan, VirusTotal, SecurityTrails APIs
• AI Analysis — OpenAI-compatible LLM generates executive summaries, attack chains, prioritized findings, and tech-specific recommendations
• Smart Pipeline — tools automatically feed data to each other (subdomains → httpx → nuclei, crawled paths → ffuf, tech detection → nuclei tags)
• 4 scan presets — quick, full, recon, web
• 5 report formats — JSON, Markdown, HTML (dark theme), SARIF (CI/CD), Web Dashboard
• Risk scoring — automatic NONE/LOW/MEDIUM/HIGH assessment
• Parallel execution — asyncio-based concurrent tool execution (3-5x faster)
• Notifications — Telegram, Discord, Slack webhooks
• Plugin system — drop-in Python plugins
• Resume — continue interrupted scans
• Incremental scanning — diff engine to track new/resolved findings

Install

git clone https://github.com/cortexc0de/argus-lite.git ~/argus-lite
cd ~/argus-lite && sudo ./install.sh

Quick Start

argus scan example.com --preset full --output html --ai
argus dashboard

Stats

• 411 tests, 84% coverage
• 75+ source files
• Built with SDD/TDD methodology