How-toCBOMNIST IR 8547

Finding weak cryptography with a CBOM

"Weak cryptography" covers three different problems: algorithms that are broken today, algorithms that are fine today but quantum-vulnerable, and symmetric key sizes. Each has its own source and timeline. This guide separates them and shows which CycloneDX CBOM fields to filter for each.

8 min readReviewed 8 November 20268 sources

A CBOM (cryptographic bill of materials) is only useful if you can ask it questions. The most common question is also the vaguest: "where is our weak cryptography?" This guide breaks that question into three buckets that need different answers, then turns each into concrete filters over the CycloneDX fields where the evidence sits. Field names below are from the CycloneDX 1.7 JSON schema, the current release[7]; most also exist in 1.6.

Three buckets, three timelines

A finding that is broken today needs an incident-style fix. A finding that is quantum-vulnerable needs a migration plan. A symmetric key-size finding may need nothing at all.

  • Bucket A: broken or disallowed today. Weak against ordinary computers. Fix now, whatever your post-quantum (PQC) plans.
  • Bucket B: fine today, quantum-vulnerable. RSA, elliptic-curve and finite-field Diffie-Hellman public-key cryptography at approved sizes. Safe against classical attack, but on NIST's deprecation clock.
  • Bucket C: symmetric and hash sizing. AES and SHA-2 are far less exposed to quantum attack. Check sizes, but do not treat AES-128 as broken.

Bucket A: classically broken or disallowed

Old TLS versions

RFC 8996 moved TLS 1.0 and 1.1 to Historic status in March 2021 and says each "MUST NOT be used"[1]. In a CBOM, look for protocol assets where protocolProperties.type is tls and protocolProperties.version is 1.0 or 1.1.

Short RSA and other sub-112-bit keys

NIST SP 800-131A Rev 2 (March 2019) disallows generating digital signatures with keys giving less than 112 bits of security: RSA with a modulus under 2048 bits, or ECDSA with an order under 224 bits. Verifying old signatures at those sizes is "legacy use" only[2]. Its Table 5 also disallows RSA key agreement and key transport with a modulus under 2048 bits[2].

In the CBOM, you have two places to look. On algorithm assets, filter algorithmProperties.classicalSecurityLevel < 112. On key assets, filter relatedCryptoMaterialProperties.type = public-key or private-key and check size. The size is in bits, so a 256-bit elliptic-curve key and a 1024-bit RSA key are not comparable on that number alone. Resolve algorithmRef first (in CycloneDX 1.7 it is deprecated, so also check relatedCryptographicAssets) and judge size per algorithm.

SHA-1 and MD5 signatures

SP 800-131A Rev 2 disallows SHA-1 for digital signature generation except where NIST protocol-specific guidance allows it, and allows SHA-1 signature verification for legacy use only[2]. RFC 6151 states that MD5 "is no longer acceptable where collision resistance is required such as digital signatures"[4]. For certificates, follow certificateProperties.signatureAlgorithmRef to the algorithm asset and check its name and parameterSetIdentifier. In CycloneDX 1.7 that field is marked deprecated in favour of relatedCryptographicAssets, so a good query checks both[7].

Triple DES (TDEA)

Two-key TDEA encryption is disallowed. Three-key TDEA encryption was deprecated through 31 December 2023 and is disallowed after that, unless other NIST guidance specifically allows it. Decryption remains for legacy use[2]. Query algorithm assets with primitive = block-cipher and a 3DES/TDEA name or algorithmFamily.

ECB mode

The draft SP 800-131A Rev 3 proposes that ECB mode be "disallowed for data encryption", with legacy use for decryption[3]. As of 8 November 2026, Rev 3 is still the initial public draft of October 2024; Rev 2 remains the final version. So call ECB a strong finding, but cite it as a draft proposal. The query is simple: algorithmProperties.mode = ecb.

Static RSA key exchange

TLS 1.3 (RFC 9846, which obsoletes RFC 8446) removed static RSA and Diffie-Hellman cipher suites, so every public-key exchange it offers provides forward secrecy[5]. Without forward secrecy, anyone who later obtains the server's RSA key can decrypt recorded sessions. SP 800-131A Rev 2 also disallows the PKCS#1 v1.5 padding scheme for RSA key transport after 31 December 2023[2]. Look for cipherSuites[].name or identifiers beginning TLS_RSA_WITH_, and for algorithm assets with padding = pkcs1v15 used for key transport.

Bucket B: fine today, quantum-vulnerable

NIST IR 8547, published final on 27 October 2026, sets one schedule for quantum-vulnerable public-key cryptography: deprecated after 2030, disallowed after 2035. This applies to RSA, ECDSA and EdDSA signatures, and to finite-field and elliptic-curve Diffie-Hellman and RSA key establishment, at every approved size[6]. RSA-2048 on a payments API, an ECDSA P-256 code-signing key and an X25519 key exchange all sit here. A larger RSA key does not move you out of this bucket.

The CBOM field built for this is algorithmProperties.nistQuantumSecurityLevel. A value of 0 means none of NIST's post-quantum security categories is met[7]. Filter for 0 on signature, key-agree, pke and kem primitives. The producer sets this value, so cross-check it against the algorithm name.

Two further queries make this bucket usable:

  • Certificates that outlive the deadline. Filter certificates whose signing algorithm is quantum-vulnerable and whose certificateProperties.notValidAfter is after 31 December 2030. Each one should be planned for re-issuance before 2035, or by 31 December 2031 for US federal high-value systems. See PKI rotation for PQC.
  • TLS without a hybrid group. In CycloneDX 1.7, cipherSuites[].tlsGroups records the key-exchange groups[7]. Endpoints listing only classical groups (for example x25519 or secp256r1) and no ML-KEM hybrid group are quantum-vulnerable key exchange. See hybrid PQC TLS.

Bucket C: symmetric and hash sizing

Grover's algorithm speeds up brute-force key search, which is often summarised as "halving" symmetric key strength. NIST's position in IR 8547 is more measured. Its approved symmetric primitives with at least 128 bits of classical security, including AES-128, "are believed to meet the requirements of at least Category 1 security", and AES-128 is the reference primitive for Category 1[6]. AES-128 is not a finding to fix. AES-256 (Category 5) is a reasonable choice for new long-lived data, but that is a design choice, not a NIST requirement.

What does matter: IR 8547 says NIST's few symmetric standards at the 112-bit level "will be disallowed in 2030", and SHA-1 will be disallowed in FIPS-validated modules after 2030[6]. Query classicalSecurityLevel = 112 on symmetric and hash primitives, and nistQuantumSecurityLevel on block ciphers (1 for AES-128, 3 for AES-192, 5 for AES-256).

FindingWhy it mattersCBOM field to filterSource
TLS 1.0 / 1.1"MUST NOT be used"protocolProperties.type = tls, version in 1.0, 1.1RFC 8996 §4–5[1]
RSA < 2048 / under 112-bitSignature generation and RSA key transport disallowedclassicalSecurityLevel < 112; relatedCryptoMaterialProperties.size + algorithmRefSP 800-131A Rev 2, Tables 2 and 5[2]
SHA-1 signatureGeneration disallowed; verification legacy use onlycertificateProperties.signatureAlgorithmRef / relatedCryptographicAssetsSP 800-131A Rev 2, Table 8[2]
MD5 signatureNot acceptable where collision resistance is neededAlgorithm name via certificate signature referenceRFC 6151[4]
3DES / TDEA encryptionDisallowed after 2023 (three-key); disallowed (two-key)primitive = block-cipher, name or algorithmFamily 3DESSP 800-131A Rev 2, Table 1[2]
ECB modeProposed: disallowed for data encryptionalgorithmProperties.mode = ecbSP 800-131A Rev 3 draft, Table 2[3]
Static RSA key exchangeNo forward secrecy; PKCS#1 v1.5 key transport disallowed after 2023cipherSuites[] TLS_RSA_WITH_*; padding = pkcs1v15RFC 9846 §1.3[5]; SP 800-131A Rev 2, Table 5[2]
RSA, ECDSA, EdDSA, ECDH, X25519 at any sizeDeprecated after 2030, disallowed after 2035nistQuantumSecurityLevel = 0; primitiveNIST IR 8547 §4.1.1–4.1.2[6]
Quantum-vulnerable certificate valid past 2030Plan re-issuance before the 2035 disallowancecertificateProperties.notValidAfter + signature algorithmNIST IR 8547 §4.1.1[6]
112-bit symmetric or hashTo be disallowed in 2030classicalSecurityLevel = 112NIST IR 8547 §4.1.3[6]
AES-128Category 1: not a findingparameterSetIdentifier = 128; nistQuantumSecurityLevel = 1NIST IR 8547 §4.1.3, Table 6[6]
Figure 1 · Weak-cryptography findings mapped to CycloneDX CBOM fields. Source: RFC 8996, RFC 6151, RFC 9846, NIST SP 800-131A Rev 2 and Rev 3 (draft), NIST IR 8547, CycloneDX 1.7 schema.

Prioritise by exposure, not by count

A CBOM with 4,000 findings is not a plan. CERT-In's v2.0 guidelines say quantum-safe implementation "should be guided by a thorough evaluation of threat exposure", technical feasibility and risk tolerance (§8.5.3). They also tell organisations to prohibit "weak or deprecated algorithms" in TLS configurations (§8.4.2.2)[8]. In practice, three questions sort the list:

  1. Is it reachable from the internet? An internet-facing TLS 1.0 endpoint or static-RSA suite is a bucket-A emergency. The same suite on an isolated lab box is not.
  2. How long must the data stay secret? Recorded traffic protected by quantum-vulnerable key exchange can be decrypted later. NIST expects to prioritise quantum-resistant key establishment to protect against "harvest now, decrypt later" attacks, particularly in TLS and IKE[6]. See harvest now, decrypt later.
  3. How long will migration take? Weigh data shelf-life plus migration time against your estimate of when a quantum computer arrives. See Mosca's inequality for BFSI.

Fix bucket A first, ordered by exposure. In bucket B, start with key exchange on long-lived sensitive data and long-validity certificates. Leave bucket C to normal lifecycle work unless you find 112-bit primitives. For checking that a CBOM can answer these queries at all, see validating a CBOM.

How KavachQ helps

KavachQ Discover runs these checks for you:

  • Assess ranks the top issues across the estate, from classically broken algorithms and weak certificate signatures to keys Shor's algorithm breaks.
  • Policy-as-code sets your own baseline, with tenant overrides, and lists every violation.
  • Each algorithm finding carries a CWE mapping and an algorithm-level FIPS 140-3 approval status (not a module-validation check). Software components are matched to CVEs and ranked with CISA KEV and EPSS, and a harvest-now-decrypt-later view shows which exposures to fix first.

Try it on a public domain with the free scan.

Frequently asked questions

Is AES-128 safe against quantum computers?

NIST treats AES-128 as meeting post-quantum security Category 1, and in IR 8547 it says approved symmetric primitives with at least 128 bits of classical security are believed to meet at least Category 1. Grover's algorithm does not make AES-128 a finding to fix. AES-256 is a reasonable choice for new long-lived data, but NIST does not require it.

Is RSA-2048 still allowed?

Yes, today. NIST SP 800-131A Rev 2 treats RSA at 2048 bits or more as acceptable. NIST IR 8547 (final, October 2026) deprecates all quantum-vulnerable public-key algorithms, RSA included, after 2030 and disallows them after 2035, whatever the key size.

How do I find TLS 1.0 and 1.1 in a CycloneDX CBOM?

Filter components whose cryptoProperties.assetType is protocol, whose protocolProperties.type is tls and whose protocolProperties.version is 1.0 or 1.1. RFC 8996 says both versions must not be used.

Is SP 800-131A Rev 3 final?

No. As of 8 November 2026, NIST SP 800-131A Rev 3 is still the initial public draft published in October 2024. Rev 2 (March 2019) is the final version. Proposals in Rev 3, such as disallowing ECB mode for data encryption, should be cited as draft.

What does nistQuantumSecurityLevel 0 mean in a CBOM?

In the CycloneDX schema, a nistQuantumSecurityLevel of 0 means the algorithm meets none of NIST's post-quantum security categories. RSA, ECDSA and ECDH will typically be recorded as 0. The value is set by whoever produced the CBOM, so cross-check it against the algorithm name.

Sources

  1. IETF, RFC 8996: Deprecating TLS 1.0 and TLS 1.1 (March 2021) — www.rfc-editor.org/rfc/rfc8996.html
  2. NIST, SP 800-131A Rev 2: Transitioning the Use of Cryptographic Algorithms and Key Lengths (March 2019) — doi.org/10.6028/NIST.SP.800-131Ar2
  3. NIST, SP 800-131A Rev 3 initial public draft (October 2024) — csrc.nist.gov/pubs/sp/800/131/a/r3/ipd
  4. IETF, RFC 6151: Updated Security Considerations for the MD5 Message-Digest and the HMAC-MD5 Algorithms (March 2011) — www.rfc-editor.org/rfc/rfc6151.html
  5. IETF, RFC 9846: The Transport Layer Security (TLS) Protocol Version 1.3 (July 2026) — www.rfc-editor.org/rfc/rfc9846.html
  6. NIST, IR 8547: Transition to Post-Quantum Cryptography Standards (October 2026) — csrc.nist.gov/pubs/ir/8547/final
  7. OWASP CycloneDX, Specification 1.7 JSON schema (current) — cyclonedx.org/docs/1.7/json
  8. CERT-In, Technical Guidelines on SBOM, QBOM & CBOM, AIBOM and HBOM v2.0 (July 2025) — cert-in.org.in/PDF/TechnicalGuidelines-on-SBOM,QBOM&CBOM,AIBOM_and_HBOM_ver2.0.pdf

Sources checked 8 November 2026. Standards and regulations change; always read the issuing body's current publication. This guide is general information, not legal advice.

CBOM, signed

See what your estate actually runs

Start with a free scan of a public domain, then scope a pilot that produces a signed CBOM for a real slice of your estate.