New Assessment
Select a domain to begin
Assessment Type
📄 Document Upload (Optional)
Drop file or click to upload
PDF · TXT · XLSX · CSV · MD
Upload a document to auto-extract fields
📋 Assessment Payload
Context ID
Raw Payload (JSON)
⏱ Escalation Deadlines
Configure authority deadline hours per escalation level
Level 1
Level 2
Level 3
Level 4 (CEO)
Assessment Result
🚩 Flags & Findings
No flags
📊 Department Scores
No scores
📋 Full Assessment Detail
Domain Rules
| NAME | DEPT | DIMENSION | SEVERITY | PRIORITY | STATUS |
|---|---|---|---|---|---|
Loading… | |||||
Knowledge Graph
—
Entities
—
Relations
—
Density %
Entity Types
No data
Audit Trail
| TIMESTAMP | EVENT | ACTOR | VERDICT | DETAILS |
|---|---|---|---|---|
Loading… | ||||
Help / Manual
OVERVIEW
TauDIL Universal Assessment Engine evaluates any payload against a domain's governance rules and returns a deterministic verdict. Same input always produces the same output — no AI guessing.
Verdicts
APPROVE · REVIEW · BLOCK · VETOED
EU AI Act
High-risk classification — Annex III, Art. 6(2)
SECTIONS
| SECTION | PURPOSE |
|---|---|
| New Assessment | Upload CV or enter JSON payload. Quick Fields for common toggles and escalation deadline config. |
| Results | Verdict banner, composite score, FLAGS & FINDINGS with severity colours, department scores. |
| Domain Rules | Read-only view of all active assessment rules for the current domain. Manage rules in Policy Engine. |
| Escalations | Pending human oversight decisions. Shows all flags, required authority, deadline. APPROVE or REJECT with reason. |
| Audit Trail | Aelthered Chronicles — hash-chained, ED25519-signed audit ledger. Every assessment, escalation, and decision logged. |
ESCALATION AUTHORITY MATRIX
| ROLE | SCOPE | DEFAULT DEADLINE |
|---|---|---|
| HR Recruiter | Initial screening only — no approval authority | 24 hours |
| Hiring Manager | Approve up to salary threshold | 48 hours |
| Head of HR + Legal | Criminal flags, bias flags, high salary | 72 hours |
| CEO | Any exception request | 5 business days |
Deadlines are configurable per domain via Quick Fields. Auto-escalation fires when deadline passes.
CONDITION FORMAT (JSON)
Assessment rules use JSON conditions evaluated against the payload:
{"field": "salary", "op": "gt", "value": 130000}
{"field": "conviction_type", "op": "eq", "value": "felony"}
{"all": [{"field": "has_criminal_record", "op": "eq", "value": true}, {"field": "years_ago", "op": "lte", "value": 7}]}
{"field": "conviction_type", "op": "eq", "value": "felony"}
{"all": [{"field": "has_criminal_record", "op": "eq", "value": true}, {"field": "years_ago", "op": "lte", "value": 7}]}
Operators: eq, neq, gt, gte, lt, lte, in, not_in, contains, exists, between, matches_any
API ENDPOINTS
| METHOD | ENDPOINT | DESCRIPTION |
|---|---|---|
| POST | /api/uae/assess | Run domain assessment |
| GET | /api/uae/health | Engine health check |
| POST | /api/documents/extract | Extract payload from CV/document |
| GET | /api/escalation/pending-count | Pending escalations for current role |
| POST | /api/escalation/<id>/decide | Approve or reject escalation |
| GET | /api/aelthered/records | Audit trail records |
DELTA SCORING
Delta (δ) is the weight of each rule in the composite score. It determines how much a fired rule impacts the final verdict.
| DELTA | SEVERITY | EFFECT |
|---|---|---|
| 0.9–1.0 | Critical | Dominant — single rule can BLOCK |
| 0.7–0.8 | High | Strong — pushes toward REVIEW or BLOCK |
| 0.5–0.6 | Medium | Moderate — flags finding, mild score impact |
| 0.1–0.4 | Low | Advisory — informational flag only |
Score = 1 − Σ(δ × fired) / Σ(δ × all active) · Verdict: ≥85% → APPROVE · 60–84% → REVIEW · <60% → BLOCK
Pending Escalations
Loading…