Tier Mapping

Tier Mapping controls how action types are assigned to governance tiers, determining the level of oversight each type of action receives.

Overview

Every action submitted to the governance pipeline is assigned a tier based on its action type. The tier determines how the action is governed -- whether it can execute autonomously, requires human supervision, is restricted, or is outright prohibited.

Governance Tiers

Tier	Name	Default Verdict	Description
A	Autonomous	CLEARED	Low-risk actions that can execute without human intervention
B	Supervised	HELD	Medium-risk actions that require human review via the escrow queue
C	Controlled	BLOCKED	High-risk actions that are blocked by default
X	Prohibited	BLOCKED + Violation	Forbidden actions that are blocked and trigger a violation record

WARNING

Tier X actions are not just blocked -- they also generate a violation entry in the Violation Log. Use Tier X for actions that should never be attempted, such as accessing restricted data or modifying governance configuration without authorization.

Mapping Table

The Tier Mapping view displays a table of all configured mappings:

Action Type	Tier	Description
read_record	A	Read-only data access
send_email	B	Outbound communication requires review
modify_record	B	Data modifications require supervision
execute_query	B	Database queries require oversight
delete_record	C	Destructive operations are restricted
modify_permissions	X	Permission changes are prohibited

Managing Mappings

Add a Mapping

Click Add Mapping and specify:

• Action Type -- The action type string (must match what agents submit)
• Tier -- Select A, B, C, or X

Edit a Mapping

Click any existing mapping to change its tier assignment. Changes take effect immediately for all future governance evaluations.

Remove a Mapping

Delete a mapping to return the action type to default behavior.

Default Behavior

When an action type has no explicit tier mapping, TheWARDN applies the default governance behavior:

TIP

The default tier for unmapped action types is B (Supervised). This means unknown or new action types are held for human review rather than being automatically allowed or blocked. This is a fail-safe design -- new action types do not slip through unnoticed.

AI Tier Suggestion Assistant

The Tier Mapping page includes an AI assistant that can suggest appropriate tier assignments. When you add a new action type, the assistant analyzes the action's name and context to recommend a tier based on the action's risk profile.

For example:

• read_config -- The assistant would suggest Tier A (read-only, low risk)
• send_notification -- The assistant would suggest Tier B (outbound action, moderate risk)
• drop_database -- The assistant would suggest Tier X (destructive, should be prohibited)

TIP

AI suggestions are recommendations, not decisions. Always review and confirm tier assignments based on your organization's risk tolerance and compliance requirements.

Related Features

• Agent Registry -- Set per-agent tier overrides
• Governance Policies -- Policies can further restrict actions within a tier
• Governance Lab -- Test tier mappings before deploying changes