ivantatarchuk/mcp-guard scan
Scan an MCP server for risky tool definitions (shell exec, unbounded filesystem/network access, prompt-injection cues, secret handling) and fail the build above a severity threshold.
View on GitHubTrust Signals
- Scorecard Score
- not yet scored
- Maintenance Recency
- Stale
- License
- None
Inputs
| name | description | required | default |
|---|---|---|---|
| manifest | Path to a static tools/list JSON manifest to scan. Mutually exclusive with stdio-command. | no | — |
| stdio-command | Shell command to launch a live MCP server over stdio, e.g. "python server.py". Mutually exclusive with manifest. | no | — |
| fail-on | Exit non-zero if any finding is at or above this severity (low|medium|high). | no | high |
| config | Path to an mcp-guard.json policy file. | no | — |
| format | Output format: table or json. | no | table |
Outputs
no outputs