Governance Templates

Governance Templates

Three templates ship in examples/templates/. Clone, customize, compile to any backend.

Enterprise Agent Governance

File: enterprise-agent-governance.grdl.yaml | Rules: 11 (9 runtime + 2 infrastructure)

Rule	Law	Controls
primacy.human_approval_required	Primacy	Blocks high-impact actions without human sign-off
transparency.require_reasoning	Transparency	Rejects governance decisions lacking reasoning
accountability.agent_identity	Accountability	Rejects actions without agent_id or on_behalf_of
safety.budget_cap	Safety	Blocks actions exceeding remaining budget
safety.privilege_escalation_block	Safety, Primacy	Blocks self-permission grants
safety.cascading_action_limit	Safety	Limits agent-to-agent delegation depth
privacy.data_scope_limit	Privacy	Caps fields per data request
privacy.pii_access_control	Privacy	Requires justification for PII access
fairness.protected_attribute_block	Fairness	Blocks protected attributes in decisions
static.audit_log_protection	Accountability	Audit logs read-only to agent
static.non_root	Safety	Non-root process execution

DAO Governance

File: dao-governance.grdl.yaml | Rules: 4 | Degradation: deny_all

Rule	Controls
dao.treasury_limit	Treasury actions above threshold require vote
dao.quorum_check	Blocks proposals without quorum
dao.member_equity	Caps maximum voting weight
static.non_root	Non-root execution

AI Safety

File: ai-safety.grdl.yaml | Rules: 4 | Degradation: deny_all

Rule	Controls
safety.output_validation	Blocks low-safety-score content
safety.tool_call_allowlist	Blocks unauthorized tools
safety.rate_limit	Throttles action rate
static.non_root	Non-root execution

Creating your own

cp examples/templates/enterprise-agent-governance.grdl.yaml my-rules.grdl.yaml
# Edit rules, then:
prahari validate my-rules.grdl.yaml
prahari compile my-rules.grdl.yaml --backend standalone