On this page
- RNX-110 · Technical Discoverability & Ingestion Efficiency
- RNX-120 · Semantic Architecture & Entity Clarity
- RNX-130 · Content Syntactics & Token Efficiency
- RNX-140 · Conflict Resolution & Data Determinism
- RNX-150 · Headless Perception & Rendering Guardrails
- RNX-160 · Geospatial & Neighborhood Anchoring
- RNX-170 · Negative Entity Logic & Safety
- RNX-180 · Multimodal Synchronization
- RNX-190 · Hallucination & Recency Guardrails
RNX code library · 88 controls
RNX-100Perception Standard
Read what these checks mean, why they matter and which evidence they concern. Each ID below keeps its canonical place in the RNX hierarchy.
These are free learning summaries, not full technical tests or an industry certification. Specialized checks apply only to the relevant feature. Experimental labels are preserved from the canonical source.
RNX-110 · Technical Discoverability & Ingestion Efficiency
47 controls. Check that the website's name leads to the intended server.
RNX-120 · Semantic Architecture & Entity Clarity
8 controls. Check that structured data can be read without syntax errors.
RNX-130 · Content Syntactics & Token Efficiency
7 controls. Give the page a clear main subject heading.
RNX-140 · Conflict Resolution & Data Determinism
RNX-141 · Universal Identity Consistency
RNX-141.T001
Current identity consistency
core readiness
Resolve verified conflicts in current business identity within the reviewed sources.
Why it matters. Different names, locations or contact facts can confuse customers.
Illustration — fictional. The contact page lists an old phone number while the approved current business profile lists the replacement number.
When it applies. Inspected target: current published identity fact set. Use this check only when that object or feature exists in the review scope; provider-specific checks require that provider's integration. Missing evidence remains unknown.
RNX-142 · Temporal Truth Labelling
RNX-142.T001
Honest modified date
core readiness
Use modified dates for real substantive updates.
Why it matters. Changing a date alone should not make old information appear newly reviewed.
Illustration — fictional. A page labels itself modified today on every visit, even when its content has not changed.
When it applies. Inspected target: published modification timestamp. Use this check only when that object or feature exists in the review scope; provider-specific checks require that provider's integration. Missing evidence remains unknown.
RNX-143 · Verifiable Freshness Provenance
RNX-143.T001
Freshness provenance trace
Experimental — not an established ranking requirement
Keep evidence for claimed observation or update times.
Why it matters. Readers and reviewers need to know what the freshness claim is based on.
Illustration — fictional. A last verified today badge has no source record or actual verification event.
When it applies. Inspected target: fact freshness provenance chain. Use this check only when that object or feature exists in the review scope; provider-specific checks require that provider's integration. Missing evidence remains unknown.
RNX-150 · Headless Perception & Rendering Guardrails
7 controls. Check that the intended reader or crawler can load the main content.
RNX-160 · Geospatial & Neighborhood Anchoring
RNX-161 · Landmark Proximity Mapping
RNX-161.T001
Landmark claim accuracy
core readiness
Check claims about nearby landmarks and locations.
Why it matters. Being near a place must not imply an office that does not exist.
Illustration — fictional. A business advertises itself as beside the airport while its actual storefront is several towns away.
When it applies. Inspected target: landmark proximity statement. Use this check only when that object or feature exists in the review scope; provider-specific checks require that provider's integration. Missing evidence remains unknown.
RNX-170 · Negative Entity Logic & Safety
RNX-171 · Boundary Definition
RNX-171.T001
Material offering restriction
core readiness
Make important service restrictions clear before a customer commits.
Why it matters. Eligibility, service limits and exclusions affect whether the offer fits.
Illustration — fictional. Wholesale-only eligibility first appears after a retail buyer submits an order.
When it applies. Inspected target: material offer restriction. Use this check only when that object or feature exists in the review scope; provider-specific checks require that provider's integration. Missing evidence remains unknown.
RNX-172 · Ambiguity Suppression Signals
RNX-172.T001
Restriction consistency
Experimental — not an established ranking requirement
Keep those restrictions consistent across current sources.
Why it matters. A booking page should not contradict the service page's limits.
Illustration — fictional. The service page says adults only while its booking feed offers appointments for children.
When it applies. Inspected target: cross-surface restriction statement set. Use this check only when that object or feature exists in the review scope; provider-specific checks require that provider's integration. Missing evidence remains unknown.
RNX-175 · Enterprise-Safe Language Calibration
RNX-175.T001
Ambiguous safety wording review
Experimental — not an established ranking requirement
Avoid misleading safety language and unsupported assurances.
Why it matters. Gentle wording must not hide a material risk or limitation.
Illustration — fictional. A serious medical limitation is replaced by euphemistic wording intended to bypass a filter.
When it applies. Inspected target: safety-language claim review. Use this check only when that object or feature exists in the review scope; provider-specific checks require that provider's integration. Missing evidence remains unknown.
RNX-180 · Multimodal Synchronization
9 controls. Give informative images useful text alternatives.
RNX-190 · Hallucination & Recency Guardrails
RNX-191 · Hallucination Stress-Testing
RNX-191.T001
Observed AI factual error test
Experimental — not an established ranking requirement
When testing AI factual errors, retain enough detail to reproduce the judgment.
Why it matters. One wrong answer does not prove every answer will be wrong.
Illustration — fictional. A captured answer claims the shop is open all night despite its documented daytime hours.
When it applies. Inspected target: factual claim. Use this check only when that object or feature exists in the review scope; provider-specific checks require that provider's integration. Missing evidence remains unknown.
RNX-192 · Retrieval Integrity Scoring
RNX-192.T001
Sample-denominator integrity
Experimental — not an established ranking requirement
Keep the sample size, missing runs and question group with reported rates.
Why it matters. A success percentage can mislead when failed or missing observations vanish.
Illustration — fictional. Unfavorable runs omitted from denominator.
When it applies. Inspected target: sampling denominator. Use this check only when that object or feature exists in the review scope; provider-specific checks require that provider's integration. Missing evidence remains unknown.
RNX-195 · Knowledge Cutoff Bridging
RNX-195.T001
Current fact accessibility
Experimental — not an established ranking requirement
Make current facts accessible and honestly dated.
Why it matters. Publishing an update does not automatically replace an AI model's older stored knowledge.
Illustration — fictional. Expired hours labeled reviewed today.
When it applies. Inspected target: fact snapshot. Use this check only when that object or feature exists in the review scope; provider-specific checks require that provider's integration. Missing evidence remains unknown.