TAAS Routing & Audit

How requests are mapped to the right niche and model, the empirical evidence behind those choices, and the compliance side-flow that keeps every tool routing through the model router. The audit is the product.

Snapshot 2026-10-14 05:00 UTC · auto-refreshed

98.87%
Router Compliance
159062
Niche Resolutions
3682168
Routed Calls

Empirical Model Evidence

Head-to-head results that drive niche scores — not vendor claims, actual outputs verified by TAAS.

Code-refactor routing: which model produces valid Python · coding / code_refactor · 2026-05-20

ModelResultDetail
claude-sonnet-4-6PASS2338 tokens, end_turn, valid Python
meta-llama/Llama-3.3-70B-Instruct-TurboFAILinvalid Python (unterminated string)
gpt-4oFAILinvalid Python (unterminated string)
claude-haiku-4-5FAILinvalid Python (unterminated string)

→ Re-scored the coding niche: claude-sonnet-4-6 ranked #1; the three failing models demoted. code_refactor added as a first-class scored niche.

Smart Niche Resolution

Unknown or newly-phrased niches resolved to the nearest scored niche (semantic + lexical), or flagged genuinely new and queued for benchmarking — instead of falling back to a blind default.

RequestedResolved → nicheMethodConfidenceWhen
research_synthesisresearch (research)lexical0.8172026-10-14 04:49
legal_analysisNEW — queued for scoringnew_niche—2026-10-14 04:47
data_classificationNEW — queued for scoringnew_niche—2026-10-14 04:44

Model-Router Compliance Side-Flow

Every tool must route LLM/AI calls through the model router. The nightly audit detects drift, auto-fixes via the dev system, escalates what it can't fix, and quarantines as a last resort.

RunScannedCompliantViolatorsFixedEscalatedWhen
v5_audit_20261014_0300291103710912125002026-10-14 03:00
v5_audit_20261013_0300201103710912125002026-10-13 03:00
v5_audit_20261012_0300221103710912125002026-10-12 03:00
v5_audit_20261011_0300181103710912125002026-10-11 03:00
v5_audit_20261010_0300181103710912125002026-10-10 03:00
v5_audit_20261009_0300271103710912125002026-10-09 03:00
v5_audit_20261008_0306441103810913125002026-10-08 03:06
v5_audit_20261007_0300211103810913125002026-10-07 03:00
v5_audit_20261005_0300221103510910125002026-10-05 03:00
v5_audit_20261004_0300291103510910125002026-10-04 03:00
v5_audit_20261003_0300181103510910125002026-10-03 03:00
v5_audit_20261002_0300411103710912125002026-10-02 03:00
v5_audit_20261001_0300121103610911125002026-10-01 03:00
v5_audit_20260930_0300211103510910125002026-09-30 03:00

Auto-Fix Attempts

ToolOutcomeViolationsAttemptsWhen
tools.seed_data_classification_niche_register_v1needs_escalationmodel_literal32026-09-29 03:00
tools.failure_triage_data_classification_v1needs_escalationmodel_literal32026-09-27 03:01
tools.seed_data_classification_niche_v2needs_escalationmodel_literal32026-09-26 03:01
tools.register_data_classification_niche_v1needs_escalationmodel_literal32026-09-24 03:01
pm.decompose_programneeds_escalationmodel_literal32026-09-23 03:00
uda.select_framework_v1needs_escalationmodel_literal32026-09-22 03:03
uda.generated_d9a4bac72f8eneeds_escalationmodel_literal32026-09-22 03:03
uda.generated_4334a904a0fdneeds_escalationmodel_literal32026-09-22 03:03
tools.seed_data_classification_niche_v1needs_escalationmodel_literal32026-09-22 03:02
subscription.credential_pool_v1needs_escalationrest_openai32026-09-22 03:02
stub.threed.terrain_reliefneeds_escalationrest_fal32026-09-22 03:02
uda.select_framework_v1needs_escalationmodel_literal32026-09-22 03:02
uda.generated_d9a4bac72f8eneeds_escalationmodel_literal32026-09-22 03:01
uda.generated_4334a904a0fdneeds_escalationmodel_literal32026-09-22 03:01
uda.code_generation_v1needs_escalationmodel_literal32026-09-22 03:01
tools.seed_data_classification_niche_v1needs_escalationmodel_literal32026-09-22 03:01
subscription.credential_pool_v1needs_escalationrest_openai32026-09-22 03:00
stub.threed.terrain_reliefneeds_escalationrest_fal32026-09-22 03:00
pm.decompose_programneeds_escalationmodel_literal32026-09-22 03:00
uda.code_generation_v1needs_escalationmodel_literal32026-09-21 03:04