{
  "markdown": "# GuardianShield\n\n<!-- mcp-name: io.github.sparkvibe-io/GuardianShield -->\n\n[![PyPI version](https://img.shields.io/pypi/v/guardianshield?v=1)](https://pypi.org/project/guardianshield/)\n[![Python](https://img.shields.io/pypi/pyversions/guardianshield?v=1)](https://pypi.org/project/guardianshield/)\n[![License](https://img.shields.io/github/license/sparkvibe-io/GuardianShield)](https://github.com/sparkvibe-io/GuardianShield/blob/main/LICENSE)\n[![Tests](https://img.shields.io/badge/tests-1883%20passing-brightgreen)]()\n\nUniversal AI security layer — an open-source MCP server for code scanning, PII detection, prompt injection defense, secret detection, dependency auditing, and audit logging.\n\n**Zero dependencies** · **27 MCP tools** · **5 safety profiles** · **108+ detection patterns**\n\n## Features\n\n- **Code Vulnerability Scanning** — SQL injection, XSS, command injection, path traversal with CWE IDs and auto-fix remediation\n- **Cross-line Data Flow Analysis** — DeepEngine tracks tainted data from sources to sinks across multiple lines using AST-based taint propagation (Python) and regex (JS/TS)\n- **Dependency Security** — Version-aware CVE matching against OSV.dev for PyPI, npm, Go, and Packagist ecosystems\n- **Manifest Parsing** — Auto-detects 11 formats (requirements.txt, package.json, yarn.lock, go.mod, composer.json, and more)\n- **Prompt Injection Defense** — 9+ detection patterns for instruction override, role hijacking, ChatML injection\n- **PII Detection** — Email, SSN, credit card, phone, IP — with automatic redaction in findings\n- **Secret Detection** — AWS keys, GitHub tokens, Stripe keys, JWTs, passwords, connection strings\n- **Safety Profiles** — 5 built-in profiles (general, education, healthcare, finance, children)\n- **Audit Logging** — SQLite-backed scan history with finding retrieval and filtering\n\n## Install\n\n```bash\npip install guardianshield\n```\n\n## Quick Start\n\n```bash\n# Register with Claude Code\nclaude mcp add guardianshield -- guardianshield-mcp\n\n# Or run directly\nguardianshield-mcp\n```\n\n## Editor Integration\n\n```bash\n# Claude Code\nclaude mcp add guardianshield -- guardianshield-mcp\n\n# VS Code (.vscode/mcp.json)\n{\"servers\": {\"guardianshield\": {\"type\": \"stdio\", \"command\": \"guardianshield-mcp\"}}}\n\n# Cursor (.cursor/mcp.json)\n{\"mcpServers\": {\"guardianshield\": {\"command\": \"guardianshield-mcp\"}}}\n\n# Claude Desktop (claude_desktop_config.json)\n{\"mcpServers\": {\"guardianshield\": {\"command\": \"guardianshield-mcp\"}}}\n```\n\n## MCP Tools\n\n### Scanning\n\n| Tool | Description |\n|------|-------------|\n| `scan_code` | Scan source code for vulnerabilities and hardcoded secrets |\n| `scan_file` | Scan a single file (auto-detects language from extension) |\n| `scan_directory` | Recursively scan a directory with filtering and progress streaming |\n| `scan_input` | Check user/agent input for prompt injection attempts |\n| `scan_output` | Check AI output for PII leaks and content violations |\n| `check_secrets` | Detect hardcoded secrets and credentials |\n| `scan_files` | Scan multiple files in one call |\n| `scan_diff` | Parse unified diff and scan only added lines |\n\n### Dependency Security\n\n| Tool | Description |\n|------|-------------|\n| `check_dependencies` | Check packages for known CVEs via OSV.dev (PyPI, npm, Go, Packagist) |\n| `sync_vulnerabilities` | Sync the local OSV vulnerability database |\n| `parse_manifest` | Parse any supported manifest file (11 formats) into dependency objects |\n| `scan_dependencies` | Scan a directory for manifest files and check all deps for vulnerabilities |\n\n### False Positive Management\n\n| Tool | Description |\n|------|-------------|\n| `mark_false_positive` | Mark a finding as false positive (flags future matches) |\n| `list_false_positives` | List active false positive records with optional filter |\n| `unmark_false_positive` | Remove a false positive record by fingerprint |\n\n### Engine Management\n\n| Tool | Description |\n|------|-------------|\n| `list_engines` | List available analysis engines with capabilities |\n| `set_engine` | Set active analysis engines for code scanning |\n\nThree engines ship built-in: `regex` (line-by-line pattern matching, enabled by default), `deep` (cross-line taint tracking), and `semantic` (structure-aware confidence adjustment).\n\n### CI & Developer Workflow\n\n| Tool | Description |\n|------|-------------|\n| `export_sarif` | Export findings as SARIF 2.1.0 JSON for GitHub Code Scanning and CI |\n| `save_baseline` | Save current findings as a baseline for delta scanning |\n| `scan_with_baseline` | Scan code and report only new findings vs. baseline |\n| `check_quality_gate` | Evaluate findings against severity thresholds (pass/fail/warn) |\n| `scan_files` | Scan multiple files in one call |\n| `scan_diff` | Parse unified diff and scan only added lines |\n\n### Configuration & Utilities\n\n| Tool | Description |\n|------|-------------|\n| `get_profile` | Get current safety profile configuration |\n| `set_profile` | Switch safety profile (general, education, healthcare, finance, children) |\n| `test_pattern` | Test a regex pattern against sample code for custom pattern development |\n| `audit_log` | Query the security audit log |\n| `get_findings` | Retrieve past findings with filters |\n| `shield_status` | Get health, configuration, and OSV cache statistics |\n\n## Configuration\n\nSet environment variables to customize behavior:\n\n| Variable | Description | Default |\n|----------|-------------|---------|\n| `GUARDIANSHIELD_PROFILE` | Default safety profile | `general` |\n| `GUARDIANSHIELD_AUDIT_PATH` | Path to SQLite audit database | `~/.guardianshield/audit.db` |\n| `GUARDIANSHIELD_DEBUG` | Enable debug logging (`1`) | disabled |\n\n## Documentation\n\nFull documentation: [sparkvibe-io.github.io/GuardianShield](https://sparkvibe-io.github.io/GuardianShield/)\n\n## License\n\nApache 2.0\n",
  "bytes": 5808,
  "sha": "1650a96742688b781c67a56924efee50e8bd4eb11349a84d9213c6282af34fad",
  "repo_slug": "sparkvibe-io/guardianshield",
  "fonte": "repo",
  "truncated": false,
  "api": "https://agentalog.com/api/listings/mcp_io_github_sparkvibe_io_guardianshield_66208e79/readme"
}