ai-coding-shield 1.0.0

Security auditing tool for AI development workflows, rules, skills, and MCPs
Documentation
# Changelog

All notable changes to this project will be documented in this file.

The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.1.0/),
and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0.html).

## [0.1.0] - 2024-03-20

### Added
- **MCP Scanner**: New module to audit Model Context Protocol servers.
  - Registry verification against trusted authors.
  - Risk detection (root access, promiscuous tools, network binding).
  - Configurable trust settings.
- **Threat Catalog**: YAML-based catalog with 40+ threat rules across 6 categories:
  - AI Security (`AI_RISK`).
  - Web Security (`WEB_RISK`).
  - Cryptography Risks (`CRYPTO_RISK`).
  - Secrets & Tokens (`SECRET_RISK`).
  - Prompt Security (`PROMPT_RISK`).
  - Infrastructure & Obfuscation.
- **Reporting**:
  - Rich terminal output with colors and severity badges.
  - JSON export for CI/CD.
  - HTML report generation with interactive UI.
- **CLI**:
  - `audit` command with filtering and output formatting.
  - `config` command to manage trusted entities.
  - `list` command to view the threat catalog.
  - `info` command for detailed threat descriptions.
- **CI/CD Integration**:
  - `--ci-mode` flag for non-zero exit codes.
  - GitHub Actions, GitLab CI, and Jenkins support documented.

### Security
- Implemented robust pattern matching for threat detection.
- Added default `trusted_authors` for MCP packages (`@modelcontextprotocol`, `@anthropic-ai`).