finktech-dev/MCP Verify
Validate, test, and secure Model Context Protocol (MCP) servers with comprehensive security scanning
View on GitHubTrust Signals
- Scorecard Score
- not yet scored
- Maintenance Recency
- Stale
- License
- None
Inputs
| name | description | required | default |
|---|---|---|---|
| server | Server command to validate (e.g., "node server.js" or "python server.py") | yes | — |
| transport | Transport type (http or stdio) | no | stdio |
| format | Output format: json, sarif, html | no | sarif |
| output-dir | Output directory for reports | no | mcp-reports |
| enable-fuzzing | Enable intelligent fuzzing for chaos testing | no | false |
| enable-semantic | Enable LLM-powered semantic analysis (requires ANTHROPIC_API_KEY) | no | false |
| llm-provider | LLM provider for semantic analysis (e.g., "anthropic:claude-haiku-4-5-20251001", "ollama:llama3.2") | no | anthropic:claude-haiku-4-5-20251001 |
| rules | Security rule blocks to execute (e.g., "OWASP,MCP,A") | no | OWASP,MCP,A,B,C |
| exclude-rules | Rule blocks or IDs to exclude (e.g., "SEC-001,D") | no | — |
| min-severity | Minimum severity level to report (info, low, medium, high, critical) | no | medium |
| fail-on-issues | Fail the workflow if security issues are found | no | true |
| baseline-path | Path to baseline report for regression detection | no | — |
| fail-on-degradation | Fail if security score drops compared to baseline | no | false |
| allowed-score-drop | Allowed security score drop before failing (0-100) | no | 5 |
Outputs
| name | description |
|---|---|
| report-path | Path to the generated report file |
| security-score | Overall security score (0-100) |
| critical-issues | Number of critical security issues found |
| high-issues | Number of high severity issues found |