v0.3.1 — ClawHub Registry Crawling

What's New

ClawHub Registry Integration

New malwar crawl CLI sub-app to browse, search, and scan skills directly from the ClawHub registry — the largest community skill marketplace with 5,700+ skills.

Commands:

• malwar crawl scan <slug> — fetch a skill's SKILL.md from ClawHub and scan it for threats
• malwar crawl search <query> — search the registry by keyword
• malwar crawl list — browse available skills with pagination
• malwar crawl info <slug> — view skill details and moderation flags (VirusTotal integration)
• malwar crawl url <url> — fetch and scan any remote SKILL.md by URL (not limited to ClawHub)

Housekeeping

• Pinned mkdocs-material to v9.x to prevent MkDocs 2.0 breakage
• Added copyright header verification to CI
• Hardened security scanning: continue-on-error replaces silent || true, added gitleaks secret scanning and npm audit
• Consolidated orphaned flat docs into nested structure

Stats

• 1,504 tests passing (32 new for crawl feature)
• 82% code coverage
• 26 detection rules
• CI green across all 5 jobs

Full Changelog: v0.3.0...v0.3.1

v0.3.0 — Production Readiness & Extensibility

What's New

Extensibility

• YAML DSL for custom rules — define detection rules in YAML without writing Python (#30)
• Rule testing framework — test rules against sample files with malwar test-rules (#30)
• Plugin system — load third-party detection plugins at runtime (#38)
• ML-based risk scoring — machine learning model for anomaly-aware risk scores (#37)

Infrastructure

• PostgreSQL backend — optional PostgreSQL support alongside SQLite (#34)
• Redis caching layer — cache scan results and rule compilations (#35)
• GitLab CI template — ready-to-use .gitlab-ci.yml with Code Quality reports (#30)
• Azure DevOps template — pipeline template with SARIF and ##vso annotations (#30)

Security & Compliance

• Audit logging — immutable, append-only audit trail for all API operations (#20)
• Role-based access control (RBAC) — admin, analyst, and viewer roles (#32)
• CI security scanning — Bandit SAST, pip-audit, and CycloneDX SBOM in CI pipeline

Operations

• Scheduled scanning — background scan jobs on configurable intervals (#26)
• Multi-channel notifications — Slack, email, and webhook alerts on scan completion (#36)
• Git diff scanning — scan only changed files between commits (#30)

User Experience

• Dashboard analytics — trend charts, threat breakdown, scan volume over time (#39)
• Rich TUI — interactive terminal interface for browsing scans and findings (#40)

Stats

• 1,472 tests (16 skipped) — up from 345 in v0.2.0
• 26 detection rules — up from 19
• 30+ API endpoints — up from 16
• 82% code coverage
• 14 issues closed

Breaking Changes

None. v0.3.0 is fully backward-compatible with v0.2.x configurations and databases.

Install

pip install malwar==0.3.0

docker pull ghcr.io/ap6pack/malwar:0.3.0

Full Changelog: v0.2.1...v0.3.0

Malwar v0.2.1 — Public Launch

Malwar v0.2.1 — Public Launch

The first public release of Malwar. 793 tests. 26 detection rules. Ready for production.

Install

pip install malwar

docker pull ghcr.io/ap6pack/malwar:latest

What's Included

Everything from v0.2.0 plus public launch infrastructure:

Detection Engine

• 26 detection rules across 9 categories
• 4-layer pipeline: Rule Engine + URL Crawler + LLM Analyzer + Threat Intel
• 3 threat campaigns: ClawHavoc, SnykToxic, ShadowPkg
• 96.8% accuracy, 100% recall, 95.8% precision
• Live LLM testing validated against Claude API

SDK & Integrations

• Python SDK: from malwar import scan
• LangChain integration: MalwarGuard, MalwarScanTool
• GitHub Action: scan SKILL.md files in PRs with SARIF upload
• STIX 2.1 / TAXII 2.1 export for SIEM integration
• Webhook notifications with HMAC-SHA256 signing

API & CLI

• 24 REST API endpoints with OpenAPI docs at /api/docs
• Per-key rate limiting with usage analytics
• Campaign ingestion from JSON, CSV, STIX, HTTP feeds
• IOC feed endpoint with ETag caching

Infrastructure

• Helm chart for Kubernetes deployment
• Multi-stage Dockerfile with non-root user
• DB migration system with auto-migrate on startup
• BSL-1.1 license — free for all use, commercial hosting requires agreement

Community

• MkDocs docs site at ap6pack.github.io/malwar
• Issue templates for bugs, features, and detection rules
• CONTRIBUTING.md and SECURITY.md
• GitHub Sponsors enabled

Documentation

• Getting Started
• Python SDK
• API Reference
• Detection Rules
• Deployment

---

Copyright (c) 2026 Veritas Aequitas Holdings LLC. All rights reserved.

Malwar v0.1.0 — Initial Release

Malwar v0.1.0

Static analysis engine for detecting malware in agentic AI skill files.

This is the initial release of Malwar — a 4-layer detection pipeline purpose-built for catching malware hidden in natural language skill files (SKILL.md).

Detection Engine

• 19 detection rules across 7 categories: obfuscation, prompt injection, social engineering, suspicious commands, credential exposure, data exfiltration, known malware
• 4-layer pipeline: Rule Engine (<50ms) → URL Crawler (1-5s) → LLM Analyzer (2-10s) → Threat Intel (<100ms)
• Verdict system: MALICIOUS / SUSPICIOUS / CAUTION / CLEAN with 0-100 risk scoring
• SARIF 2.1.0 output for CI/CD integration

API & CLI

• 16 REST API endpoints — scan submission, batch scan, results retrieval, SARIF export, signature CRUD, campaigns, reports
• API key authentication and per-IP rate limiting
• CLI tool — malwar scan, malwar serve, signature management, database management
• Webhook notifications for malicious scan results

Web Dashboard

• React 19 + TypeScript + Vite + Tailwind CSS 4
• Dashboard with scan stats and severity distribution
• Scan submission, history, and detail views
• Signature management and threat campaign tracking
• Served automatically by the API server

Threat Intelligence

• ClawHavoc campaign seeded with IOCs: C2 IPs, malicious domains, payload URLs, known threat actor handles
• Signature database with exact, regex, and IOC pattern matching
• Campaign attribution linking findings back to organized threats

Infrastructure

• SQLite storage with async I/O (aiosqlite)
• Docker multi-stage build with docker-compose
• CI/CD pipeline via GitHub Actions (lint, type check, test, build)
• 354 tests, 89% code coverage
• 37 test fixtures including real-world malicious samples from ClawHub and Snyk research

Quick Start

pip install malwar-0.1.0-py3-none-any.whl
malwar db init
malwar scan SKILL.md
malwar serve  # API + Dashboard at http://localhost:8000

Documentation

• Architecture
• API Reference
• Detection Rules
• Threat Campaigns
• CLI Reference
• Deployment
• Development

---

Copyright (c) 2026 Veritas Aequitas Holdings LLC. All rights reserved.