Skip to content

Latest commit

 

History

History
 
 

README.md

Policies Example

Sample YAML governance policy files for AgentMesh. Each file provides an example configuration covering a specific security or compliance scenario.

Policies

File Description
adk-governance.yaml Tool restrictions and delegation controls for Google ADK agents
cli-security-rules.yaml Blocks dangerous shell and CLI patterns
conversation-guardian.yaml Content safety rules for conversational agents
mcp-security.yaml Security controls for MCP tool usage
pii-detection.yaml Detects and blocks personally identifiable information
prompt-injection-safety.yaml Guards against prompt injection attacks
sandbox-safety.yaml Restrictions for sandboxed agent execution
semantic-policy.yaml Semantic similarity-based policy enforcement
sql-readonly.yaml Restricts agents to read-only SQL operations
sql-safety.yaml Blocks destructive SQL patterns
sql-strict.yaml Strict SQL allowlist for production databases

How to Use

These policy files can be applied to agent workflows to enforce governance rules.

scripts/check-policy.sh --action "web_search" --tokens 1500 --policy examples/policies/sql-safety.yaml

Related

  • Quickstart - Runnable examples that use these policies