CAS Registry Number
Enable an AI agent to recognise a CAS Registry Number, assess evidence for its substance assignment, and decide whether it can support a particular identification or linking task.
Research draft, second pass
A second pass drafted this model: the structure a model of this thing needs, and what is known about it in the world. The line under this one says how the second half was obtained - researched against sources, or recalled without web access, in which case nothing here was read anywhere and every claim is a lead to verify. Unreviewed either way.
Researched by: Codex + Grok
Purpose and description
Enable an AI agent to recognise a CAS Registry Number, assess evidence for its substance assignment, and decide whether it can support a particular identification or linking task.
A CAS Registry Number is a unique numeric identifier assigned by CAS, a division of the American Chemical Society, to a specific chemical substance recorded in the CAS Registry.
It can be Recognise and normalise candidate CAS RN text while preserving its original form.; Check syntax and checksum, flagging failures without silently repairing digits.; Resolve a candidate against an identified source and capture assignment evidence.; Compare the source's substance description with the intended material or record.; Link compatible records while retaining mapping evidence and unresolved conflicts.; Prepare a correction or CAS verification request when identification remains unsettled..
Distinguishing features
A CAS RN is a numeric substance identifier with three hyphen-separated parts and a final check digit; appearance alone does not establish an assignment. [CAS REGISTRY](https://www.cas.org/cas-data/cas-registry)
Its digits do not encode chemical structure or composition, unlike a structural representation such as SMILES. [CAS FAQ](https://www.cas.org/about/faqs)
A claim that a number identifies a substance requires assignment evidence, distinguishing a CAS RN from an arbitrary supplier or local inventory code.
The identification target must be a registered substance, rather than a particular bottle, batch, product listing or document.
Matching number strings establish identifier agreement; they do not independently establish that two physical materials have matching composition or quality.
Scope
+ Observed and normalised CAS RN representations
+ Format and check-digit assessment
+ Evidence connecting the number to a CAS-assigned substance
+ Compatibility between the registered substance and a proposed identification target
+ Conflicting, historical and cross-system identifier assertions
+ Readiness for lookup, linking, correction and reuse
- Chemical structure, composition and properties as independently modelled substance facts
- Physical samples, batches, purity and analytical certificates
- Commercial products, formulations, packaging and stock records
- Toxicology, clinical suitability and handling instructions
- Regulatory authorisations and jurisdiction-specific inventory decisions
- Operation and complete contents of the CAS REGISTRY database
Characteristics
- Observed representation
- Literal source text, including separators and surrounding labels Preserves evidence before transcription or normalisation changes.
- Format assessment
- unchecked | conforming | nonconforming | ambiguous Separates recognisable CAS RN syntax from damaged or mislabelled input.
- Check-digit assessment
- unchecked | pass | fail | not-computable Detects some transcription errors without claiming that CAS assigned the number.
- Assignment evidence
- unverified | CAS-confirmed | secondary-source-only | conflicting Distinguishes authoritative confirmation from repetition of an assertion.
- Registered substance referent
- Link to the substance description supported by the consulted record Makes explicit what the identifier is asserted to identify.
- Target match
- unassessed | supported | incompatible | insufficient-detail Prevents applying a valid identifier to an inadequately matched material.
- Mapping relationship
- exact-referent | broader | narrower | related | unresolved Prevents neighbouring identifiers or substance forms from becoming unsupported synonyms.
- Verification age
- Elapsed days since the recorded verification; unknown if never verified Supports task-specific decisions about rechecking the evidence.
- Use readiness
- lookup-only | identification-supported | review-required | rejected-for-target Connects validation and substance matching to an explicit next action.
Where this came from
wikidata · CC0 1.0
Drafted structure
Bundle to layer to finding to question, as the second pass will find it: 6 bundles · 11 layers · 11 findings · 24 questions.
Number recognition Establish whether observed text is a usable CAS RN candidate.
A transcription error can redirect substance lookup before any chemical reasoning occurs.
Representation
Preserve the observed number and justify its normalised form.
Candidate boundaries
Record the exact token, its CAS label and any formatting transformations.
- Which exact source text is asserted to be a CAS RN, and where does that assertion occur? provenance
- Can separators or whitespace be normalised without guessing, inserting or changing digits? action
Check-digit validation
Assess numerical integrity independently of assignment evidence.
Validation outcome
Record separate syntax and checksum results, including the validation rule used.
- Does the candidate satisfy the documented CAS RN group-length and separator rules? definition
- Does the computed check digit equal the supplied final digit? measurement
- If validation fails, which source must be revisited before proposing a corrected number? action
CAS assignment evidence Establish the evidence that CAS assigned the candidate to the stated substance.
A checksum-valid string and a supplier assertion are insufficient to confirm the assignment.
Assignment confirmation
Identify the record supporting the number-to-substance relationship.
Authority of assignment
Record whether evidence comes from CAS or an intermediary; CAS identifies itself as the final authority for these numbers. [CAS REGISTRY](https://www.cas.org/cas-data/cas-registry)
- Which consulted record explicitly connects this exact number to a substance description? provenance
- Is that connection confirmed by CAS, reproduced by another publisher or merely inferred from a name match? provenance
Lookup coverage
Interpret lookup outcomes within the consulted service's coverage and access limits.
Unresolved search
Keep a failed or unavailable lookup separate from evidence that a candidate has no assignment.
- Did the search return no match, an access failure or a record outside the source's available coverage? boundary
- Which further CAS lookup or verification route could settle the assignment? action
Substance reference boundaries Determine whether the assigned referent fits the substance being identified.
Correctly reading a CAS RN does not ensure that it has been attached to the correct chemical form or material.
Chemical form specificity
Compare identity-defining details without assuming universal registration rules for each form.
Form match
Record agreement or uncertainty concerning salt form, hydration, stereochemistry and isotopic specification.
- Which chemical-form details does the supporting CAS record actually specify? definition
- Does the intended target agree on those details, or is the apparent match based only on a shared name? boundary
Material granularity
Determine the level at which a CAS RN assertion applies within a material description.
Whole or constituent
Distinguish identification of the stated substance from identification of a constituent or related material.
- Does the source attach this number to the whole material, one constituent or a broader substance description? boundary
- For a polymer or variable-composition material, which descriptive limits must match before this number can be applied? boundary
Identifier reconciliation Manage competing CAS RN assertions and links into other identifier systems.
Automatic joins can propagate an incorrect substance assignment across otherwise independent records.
Cross-system matching
Qualify relationships to names, structures and other substance identifiers.
Mapping equivalence
Require evidence that a linked identifier describes the same referent at compatible specificity.
- What evidence connects this CAS RN to the proposed external identifier or structure? provenance
- Is the mapping exact, broader, narrower or merely related in chemical identity? boundary
Conflict and history
Preserve disagreements and any explicitly documented historical relationships.
Competing number claims
Avoid treating multiple numbers associated with one name as interchangeable without evidence.
- Which sources give competing numbers, and what substance descriptions and dates accompany each assertion? provenance
- Is there explicit authority evidence for a historical relationship, or could the discrepancy reflect different substances or an error? boundary
- Which links must remain unresolved until that discrepancy is settled? action
CAS RN use decisions Translate validation and assignment evidence into permitted identification operations.
The agent needs to distinguish a useful search key from evidence sufficient for a consequential record linkage.
Identification gates
Set the evidence threshold for the intended CAS RN operation.
Task-specific readiness
Record whether the number is ready for lookup, target identification or joining records.
- Is the intended operation exploratory lookup, substance identification or merging records? action
- Which unresolved assignment or form-matching issue prevents that operation from proceeding? action
- Would the operation incorrectly treat identifier agreement as evidence of product equivalence, purity or approval? boundary
Correction and reuse
Determine how an assertion may be corrected or carried into another system.
Evidence-preserving action
Keep local corrections distinguishable from CAS assignment decisions and retain conditions attached to retrieved records.
- Does the issue require correcting a local transcription, withdrawing a substance mapping or seeking CAS confirmation? action
- What source attribution and applicable access or reuse conditions must accompany the retrieved identifier evidence? action
Evidence and external alignment What the world already says about this thing, gathered so the model can be checked against it.
A model that cannot be lined up against existing standards, identifiers and practice cannot be adopted by anyone who already uses them.
Reported evidence
Findings from the breadth pass, kept separate from the structural claims.
Kinds and varieties
Reported by the breadth pass; each item needs checking against its source before it becomes normative.
- Public assigned CAS RN
- Deleted or replaced CAS RN
- Alternate CAS RN for the same substance
- CAS RN for a defined mixture or UVCB substance
- CAS RN for a polymer
- CAS RN for a biological sequence
- Which of these kinds and varieties hold for the sense of CAS Registry Number this model covers, and on what evidence? provenance
Sources
- CAS Registry - Official definition of the CAS Registry and assignment of unique CAS Registry Numbers to indexed substances.
What the second pass must settle
- What exact CAS documentation should govern syntax, leading-zero treatment and checksum implementation in this model?
- Which historical assignment relationships and status terms does CAS expose, and how should their evidence be represented?
- What registration boundaries apply to stereoisomers, salts, hydrates, isotopic forms and crystal forms in the cases this catalogue must handle?
- What descriptive evidence is sufficient to match polymers and variable-composition substances without asserting unsupported equivalence?
- Which CAS sources will agents be able to access, and what verification and reuse conditions apply to the intended deployment?