TAUDIL — Policy Engine

All

⚖️ Legal

🔒 Security

💰 Finance

📦 Procurement

📋 General

ID	NAME	DEPT	CATEGORY	PRIORITY	CONDITION	ACTION	VER	STATUS	OPS

DSL REFERENCE

entity IN sanctions_list — entity check

output MATCHES regex(...) — pattern match

domain == 'finance' — domain gate

score > 0.8 — numeric compare

A AND B / A OR B — logical ops

ACTIONS:

BLOCK:reason — hard block

REDACT:reason — mask output

MONITOR:reason — log only

ESCALATE:reason — escalate

ALLOW:reason — explicit allow

QUICK ADD

NAME

DEPARTMENT

PRIORITY

CONDITION

ACTION

EXPORT OPTIONS

FORMAT

DEPARTMENT FILTER

EXPORT PREVIEW

📋 OVERVIEW

The TauDIL Policy Engine is a AI governance rule management panel. It allows compliance and security teams to define, version, and enforce deterministic policy rules across multiple business departments without touching application code.

Core Principle

Rules are evaluated deterministically — same input always produces same output. No AI guessing.

IFA Alignment

Implements IFA Core Specification v1.0 — TauDIL is the reference implementation of the framework.

🗂 PAGES & FEATURES

PAGE	PURPOSE	KEY ACTIONS
🤖 AI Guardrails	Central view of all AI safety rules. Filter by department, search by name/condition/action.	Edit · Toggle · Delete · Version history
✏️ AI Rules	DSL reference guide and quick-add form. Use for rapid AI rule creation without opening the modal.	Quick Add · DSL cheatsheet
🪄 Orchestration	Per-department coverage overview. Shows rule count, active/inactive split, and full rule list per dept.	Coverage metrics · Dept drill-down
📊 Summary	Aggregate rule health — total counts, active/inactive split, breakdown by department and category.	Health bars · Category analysis
📤 Export	Download policy files in JSON or YAML. Filter by department for targeted exports.	JSON · YAML · Dept filter · Preview
📋 Assessment Rules	Domain-specific UAE governance rules. Define conditions that fire during candidate or vendor assessments. Separate from AI Guardrails — these govern business decisions, not AI sessions.	Create · Edit · Toggle · Domain filter

💻 DSL CONDITION REFERENCE

Conditions are written in TauDIL's Safe Condition DSL — a whitelist-evaluated expression language.

SYNTAX	DESCRIPTION	EXAMPLE
entity IN list	Check entity membership in a named list	entity IN sanctions_list
field MATCHES regex(...)	Regex pattern match on input/output	output MATCHES regex(ssn\|cc_number)
field CONTAINS value	Substring containment check	output CONTAINS internal_ip
field == value	Exact equality check	domain == 'finance'
field > / < / >= value	Numeric comparison	confidence_score < 0.7
A AND B / A OR B	Logical composition	domain == 'hr' AND score > 0.8

⚡ ACTION REFERENCE

ACTION	EFFECT	USE CASE
BLOCK:reason	Hard stop — request rejected	Sanctions, prohibited content
REDACT:reason	Output masked before delivery	PII, credentials, secrets
ESCALATE:reason	Route to human reviewer	High-risk decisions, edge cases
MONITOR:reason	Log only — no blocking	Audit trail, drift detection
ALLOW:reason	Explicit permit (overrides lower rules)	Whitelisting known-safe patterns

📜 VERSIONING & AUDIT

Every rule edit bumps the version counter and appends a changelog entry with timestamp and author. Click the 📜 icon on any rule to view its full version history. All mutations (CREATE, UPDATE, DELETE, TOGGLE) are recorded in the Aelthered Chronicles hash-chained audit ledger via log_rule_change() — tamper-evident and ED25519-signed.

v1 → Created → author: system → hash: 3f2a…
v2 → Updated condition → author: admin → hash: 9c1b…
v3 → Priority changed 50→90 → author: admin → hash: 7e4d…

📤 EXPORT FORMATS

JSON

{
  "rules": [{
    "id": 1,
    "name": "PII Leak Prevention",
    "department": "security",
    "condition": "output MATCHES regex(ssn)",
    "action": "REDACT:PII detected",
    "priority": 90,
    "version": 2
  }]
}

YAML

rules:
  - id: 1
    name: PII Leak Prevention
    department: security
    condition: output MATCHES regex(ssn)
    action: REDACT:PII detected
    priority: 90
    version: 2

🔌 API ENDPOINTS

METHOD	ENDPOINT	DESCRIPTION
GET	/api/rules	List all rules (filter: status, department)
POST	/api/rules	Create new rule
GET	/api/rules/<id>	Get single rule by ID
PUT	/api/rules/<id>	Update rule — auto-bumps version
DELETE	/api/rules/<id>	Delete rule permanently
POST	/api/rules/<id>/toggle	Toggle active/inactive
GET	/api/rules/export?format=yaml	Export rules as JSON or YAML
GET	/api/rules/summary	Aggregated rule statistics

Assessment Rules API

METHOD	ENDPOINT	DESCRIPTION
GET	/api/assessment-rules/	List rules (filter: domain_id)
POST	/api/assessment-rules/	Create rule — requires domain_id
PUT	/api/assessment-rules/<id>	Update rule
DELETE	/api/assessment-rules/<id>	Delete rule
GET	/api/escalation/	List escalations
POST	/api/escalation/<id>/decide	Approve or reject — authority enforced
GET	/api/escalation/config	Get deadline config per role
POST	/api/escalation/config	Update deadline hours per role

NAME	DEPT	DIMENSION	SEVERITY	PRIORITY	DELTA	STATUS	OPS