RFC 8949 Quiz (EN)

**Judgment point:** Separate the role of CBOR data model from nearby security functions and policy decisions. **Related keywords:** - **data item**: A single CBOR value - **tag**: Marker for additional semantics - **deterministic encoding**: Stable byte representation **Options:** - A: defining the business meaning of every map key in every protocol is not the central purpose of RFC 8949. It can be nearby work, but it is not the decision this RFC primarily supports. - B: encoding structured data in a compact binary representation with extensibility and well-defined data items is the reason this RFC belongs in the review path. It drives checks on inputs, validation, and boundaries. - C: making signed data stable without a deterministic encoding profile is outside the RFC boundary. Higher-layer policy and adjacent protocol responsibilities still need separate text.

**Judgment point:** CBOR data model is a building block. Acceptance conditions, authorization, and deployment exceptions must still be explicit. **Related keywords:** - **data item**: A single CBOR value - **tag**: Marker for additional semantics - **deterministic encoding**: Stable byte representation **Options:** - A: the schema or profile that assigns meaning to map keys, tags, and allowed value types is not decided by the RFC alone. The relying application must say what success means. - B: Skipping local acceptance policy can accept a valid object in the wrong context. That crosses the RFC boundary. - C: parse only well-formed CBOR and apply the application profile for duplicate keys, tags, and resource limits is a boundary the implementation review must preserve. Source and scope mistakes cause false acceptance. - D: An RFC citation can be necessary, but deployment configuration, keys, trust, and failure handling are not automatic.

**Judgment point:** A profile is the contract that makes use of CBOR data model consistent across implementations. **Related keywords:** - **data item**: A single CBOR value - **tag**: Marker for additional semantics - **deterministic encoding**: Stable byte representation **Options:** - A: If input bytes and context are free-form, implementations verify different values. - B: Human-readable text is useful for explanation, but it is unstable as a validation input. Locale and wording changes break it. - C: choose deterministic encoding when bytes are compared, signed, or hashed across implementations is required. Clear producer rules let verifiers evaluate the same object under the same assumptions.

**Judgment point:** Interoperability failures appear when the same wire data receives different processing meanings. **Related keywords:** - **data item**: A single CBOR value - **tag**: Marker for additional semantics - **deterministic encoding**: Stable byte representation **Options:** - A: signing one CBOR serialization while another valid serialization is accepted as equivalent shifts the validation target or meaning. Implementations can then disagree about success. - B: Limiting BCP 14 words to real requirements reduces ambiguity rather than creating it. - C: Documenting unsupported-extension behavior reduces implementation divergence.

**Judgment point:** CBOR data model needs validation against its defined inputs and scope, not just its delivery path. **Related keywords:** - **data item**: A single CBOR value - **tag**: Marker for additional semantics - **deterministic encoding**: Stable byte representation **Options:** - A: HTTPS protects transport, but it does not replace validation of CBOR data model. - B: parse only well-formed CBOR and apply the application profile for duplicate keys, tags, and resource limits is the central step. It separates well-formed input from acceptable input. - C: A filename or path can be a routing hint, but it is not enough as identity or security proof.

**Judgment point:** A specification can define processing rules while still leaving acceptance policy outside the mechanism. **Related keywords:** - **data item**: A single CBOR value - **tag**: Marker for additional semantics - **deterministic encoding**: Stable byte representation **Options:** - A: major types, data items, tags, maps, arrays, definite and indefinite lengths, and deterministic encoding choices is part of the mechanism RFC 8949 addresses directly. It is not the policy layer. - B: the schema or profile that assigns meaning to map keys, tags, and allowed value types is an application or deployment acceptance condition. - C: Well-formed syntax and algorithm processing belong to the RFC-defined side. The meaning of success still needs policy. - D: limits for nesting depth, integer size, string size, and duplicate map keys is local policy. It should be written for the deployment and threat model.

**Judgment point:** A security review looks for false-acceptance paths, not just whether the mechanism is present. **Related keywords:** - **data item**: A single CBOR value - **tag**: Marker for additional semantics - **deterministic encoding**: Stable byte representation **Options:** - A: Performance can matter, but caching does not replace a security property. - B: A recent library helps, but local policy, inputs, and threat model still need review. - C: letting different encodings of the same abstract value break signatures, hashes, or cache keys is the central failure mode. Attackers exploit this kind of context or trust shift.

**Judgment point:** Adjacent specifications are usually safer to read as layered responsibilities, not automatic replacements. **Related keywords:** - **data item**: A single CBOR value - **tag**: Marker for additional semantics - **deterministic encoding**: Stable byte representation **Options:** - A: CWT and COSE rely on CBOR, but their security processing is defined by their own profiles preserves the responsibility split. - B: Obsoletion is explicit in RFC metadata and text. Nearby concepts do not automatically replace one another. - C: CBOR data model matters in protocol composition reviews. Treating it as unrelated hides boundaries.

**Judgment point:** Review questions should show whether implementers can reproduce the same success and failure boundary. **Related keywords:** - **data item**: A single CBOR value - **tag**: Marker for additional semantics - **deterministic encoding**: Stable byte representation **Options:** - A: Naming inputs, context, validation rules, and rejection behavior directly reduces implementation differences. - B: An RFC number does not define local semantics by itself. A profile or policy must fill the gap. - C: Treating unknown values as success can accept extensions or attacker input incorrectly. - D: when a binary protocol or token profile needs compact, typed, and extensible data is a real review setting. Fixing success conditions there reduces misuse.

**Judgment point:** Validation failure is a signal to stop using that input as evidence for the decision. **Related keywords:** - **data item**: A single CBOR value - **tag**: Marker for additional semantics - **deterministic encoding**: Stable byte representation **Options:** - A: Downgrading to a warning creates fail-open behavior. Attacker-controlled input can enter the success path. - B: Treating the result as unusable for that decision is safe. If fallback exists, it needs explicit policy. - C: Authorization is important, but it does not automatically repair broken validation input.