odingard/Cerberus MCP Tool Poisoning Scan

Scan AI agent and MCP tool descriptions for hidden or malicious instructions.

View on GitHub

Trust Signals

Scorecard Score
not yet scored
Maintenance Recency
Stale
License
None
namedescriptionrequireddefault
manifestPath to a JSON tool manifest, relative to the repository root.nocerberus-tools.json
fail-onMinimum severity that fails the job (none, low, medium, or high).nohigh
reportPath for the JSON report, relative to the repository root.nocerberus-tool-scan-report.json
namedescription
scanned-countNumber of tool descriptions scanned.
poisoned-countNumber of tool descriptions with one or more findings.
high-countNumber of high-severity findings.
reportPath to the generated JSON report.