# Choosing a PQC Discovery Tool: A Practical Selection Checklist

> A practical checklist for choosing a PQC discovery tool: what each discovery method should find, accuracy tests, CBOM output, data safety, vendors and a 30-day pilot.

- Source: https://sinevis.com/choosing-a-pqc-discovery-tool/
- Publisher: Sinevis (https://sinevis.com/)
- Published: 2026-11-19
- Last updated: 2026-11-19
- Category: PQC (Post-Quantum Cryptography)

---

Many organisations start their post-quantum programme by searching for a **PQC discovery tool**, and find that the products on the first page of results promise much the same thing: find all your cryptography, flag what a quantum computer will break, produce an inventory. The demos look alike. The differences show up later, on your own network, when one tool finds the SSH keys on a jump host and another does not.

This article is a practical checklist for choosing a PQC discovery tool. It is written for CISOs, security architects and procurement teams who have a shortlist and need a fair way to test it. It covers what each discovery method should find and how to test it, the accuracy checks that catch weak tools early, what the output should look like, how the tool must handle your data, a starting list of vendors, and a 30-day pilot plan.

If you need more than discovery, such as exposure scoring, prioritisation and migration planning, our guide to [choosing a cryptographic exposure management platform](https://sinevis.com/how-to-choose-cryptographic-exposure-management-platform/) is the next step.

## Why discovery comes first

Every serious migration timeline starts with an inventory. NIST’s final [IR 8547](https://csrc.nist.gov/pubs/ir/8547/final) (October 2026) tells organisations to treat the move to post-quantum cryptography as “a current operational priority, beginning with cryptographic inventories”. The UK NCSC’s [migration timelines](https://www.ncsc.gov.uk/guidance/pqc-migration-timelines) (20 March 2025) ask organisations to carry out a full discovery exercise by 2028.

US federal guidance is explicit about tooling. OMB [Memorandum M-26-15](https://www.whitehouse.gov/wp-content/uploads/2026/06/M-26-15-Execution-of-the-Migration-to-Post-Quantum-Cryptography.pdf) (24 June 2026) says the inventory “should be created and maintained through automated tools wherever possible”, naming software composition analysis, static and dynamic application security testing and network scanners, and says the data “should populate a central Cryptographic Bill of Materials (CBOM)”. It applies to federal agencies, but it is a useful description of what a discovery tool is for.

CISA’s [strategy for automated cryptography discovery and inventory tools](https://www.cisa.gov/resources-tools/resources/strategy-migrating-automated-post-quantum-cryptography-discovery-and-inventory-tools) (15 August 2024) sets the expectation for coverage. Tools, “either individually or in combination”, provide four discovery types: network, file system, database system and software package. It is also candid about the limit: “software packages with embedded cryptography, customized applications, or certain operating systems may not be discoverable using automated tools.” Any vendor that claims otherwise is worth questioning closely.

## The selection checklist, method by method

Discovery tools are built from a handful of collection methods. Few products do all of them well, so map each candidate against the methods your estate actually needs, and turn every claim into a test you can run in a pilot.

| Method | What to ask | How to test it in a pilot |
| --- | --- | --- |
| **Network: TLS, SSH, IPsec** | Which ports and protocols are scanned by default? Is TLS behind STARTTLS covered? Does IKEv2 discovery work without establishing a tunnel? | Point it at a known range including a mail server, a VPN gateway and an SSH bastion. Compare its output with the configurations you hold. |
| **Hosts and endpoints** | Does it need its own agent, or can it use an existing EDR agent or authenticated SSH? What does it read: libraries, keystores, certificate stores, disk encryption? | Run it on a sample of Windows and Linux builds with known keystores and an old OpenSSL version installed. |
| **Databases** | Which database engines are supported, and does it connect with a read-only account? Does it check the TLS configuration of each listener, transparent data encryption (TDE) and column-level encryption, and record where the encryption keys are held? | Stand up test instances with TDE on and off and TLS set to a known protocol version and cipher suite. Check that each setting and the key location are reported correctly. |
| **Source code** | Which languages? Does it report file, line and function? Does it scan history, dependencies and secrets, or only the current branch? | Seed a test repository with calls to RSA, ECDSA, ML-KEM and a hard-coded key. See our notes on [source code and file share discovery](https://sinevis.com/cryptographic-discovery-source-code-file-shares/). |
| **Cloud KMS and cloud key vaults** | Which clouds and regions? What read-only role does each connector need? | Create test keys of different types in a non-production account and check every one appears with type and size. |
| **On-premises KMS and vaults** | Which on-premises key managers and secrets vaults have read-only connectors, for example over KMIP or the product’s own API? Does it record each key’s algorithm, size, state and rotation date without exporting key material? | Create keys of different types in a test domain or partition. Confirm each is reported with algorithm, size and state, and that the connector’s account has no permission to export keys. |
| **HSMs** | Which HSM models are named in the connector list? Is the result key metadata only? | Generate an RSA and an EC key in a test partition and confirm they are recorded without any key export. |
| **PKI and certificate authorities** | Does it read issued certificates from the CA itself, or only certificates presented on endpoints? Does it record chains? | Issue a certificate that is not bound to any listener and see whether it appears. Our article on [PKI and HSM discovery](https://sinevis.com/pki-hsm-discovery-pqc-inventory/) explains why this matters. |
| **File shares and data at rest** | Which protocols and versions? Does it identify files by content or by extension? How much of each file does it read? | Place a renamed encrypted archive and a PKCS#12 file on a test share. |
| **Passive traffic** | Is there a passive sensor for estates you cannot scan? Which protocols does it decode? | Feed it a mirror port on a test segment for a week and compare with the active results. For control networks see [passive PQC discovery in OT](https://sinevis.com/passive-pqc-discovery-ot/). |

Two passive-sensor limits are worth knowing before the pilot. A sensor only sees connections that actually happen while it watches. And under TLS 1.3, now specified in [RFC 9846](https://www.rfc-editor.org/rfc/rfc9846) (July 2026, obsoleting RFC 8446), “All handshake messages after the ServerHello are now encrypted”, so the server certificate is not visible on the wire. A passive tool can still record the negotiated cipher suite and key exchange group, but certificate details have to come from an active scan, the CA or the host.

## Accuracy tests that separate tools quickly

Coverage claims are easy to write. These checks show whether the findings can be trusted.

- **Does it recognise post-quantum and hybrid cryptography already in use?** [RFC 10024](https://www.rfc-editor.org/rfc/rfc10024) (August 2026) defines hybrid TLS 1.3 key agreement including X25519MLKEM768, which the [IANA TLS Supported Groups registry](https://www.iana.org/assignments/tls-parameters/tls-parameters.xhtml) lists as value 4588. On the SSH side, [OpenSSH 10.0](https://www.openssh.com/releasenotes.html) (April 2025) made the hybrid mlkem768x25519-sha256 its default key agreement. Stand up one endpoint of each and check the tool names them correctly rather than reporting “unknown” or counting only the classical half. Add a server configured to accept only ML-KEM groups and check the tool records it rather than logging a failed scan.
- **Does every finding say where it is used?** An algorithm without a host, port, file path, key label or line of code cannot be fixed. Check that each entry carries a location and the method that found it.
- **How are duplicates handled?** The same certificate may appear on a load balancer, in a key vault and on a back-end server. A good inventory shows one certificate with several locations. Rescan the same scope and check the count does not grow.
- **How many false positives?** Take a sample of findings, perhaps fifty, and verify each by hand. Typical errors include file-extension guesses, commented-out code and test fixtures counted as production use.
- **What happens to things it does not recognise?** An unrecognised algorithm should be recorded and flagged for review, not dropped.
- **Does it record strong cryptography as well as weak?** A tool that lists only problems cannot show migration progress.

## Output: an open CBOM and somewhere to send it

The inventory is what you are really buying, so it must be able to leave the tool. The open format to ask for is a CycloneDX Cryptography Bill of Materials. [CycloneDX 1.7](https://cyclonedx.org/news/cyclonedx-v1.7-released) was released on 21 October 2025 and is also published by Ecma as [ECMA-424](https://ecma-international.org/publications-and-standards/standards/ecma-424/) (2nd edition, December 2025). An open format means an auditor, a regulator or your next platform can read the inventory without the product that produced it, and you are not locked in.

In the pilot, ask for an export and validate it against the published schema. Check which fields are populated: algorithm, key size or parameter set, protocol, certificate details, location, first and last seen. Then check integration. Can findings be linked to assets in your CMDB, and can they raise tickets in the tracker your engineering teams already use? A finding that never reaches an owner never gets fixed. Our [tools and frameworks guide](https://sinevis.com/tools-and-frameworks-for-post-quantum-readiness/) lists open-source helpers if you want to compare a commercial export with a scripted baseline.

## Safety and data handling

A discovery tool touches HSMs, certificate authorities, key vaults and production networks. Its own security model belongs in your risk assessment.

- **Read-only collection.** Every connector should work with a read-only account, and the vendor should list the exact permissions.
- **No private key extraction.** The tool should record key metadata such as type, size and label, never key material or secret values. Ask for this in writing.
- **File and code handling.** Ask how much of each file is read, whether repository clones persist, and whether secrets found in code are stored or only located.
- **Where the data lives.** An inventory of your weak cryptography is sensitive. Confirm whether the platform runs on-premises or in your own cloud account, whether it works in air-gapped segments, and how updates reach it there.
- **Fragile estates.** In operational technology, NIST [SP 800-82 Rev. 3](https://csrc.nist.gov/pubs/sp/800/82/r3/final) says owners “should exercise extreme caution when permitting active scanning on an operational network” because active scans “may cause device instability”. Rev. 3 remains final; a [Rev. 4 initial public draft](https://csrc.nist.gov/pubs/sp/800/82/r4/ipd) was published on 21 September 2026. Ask for scan pacing controls, a list of probes that are withheld by default, and a passive option.

## PQC discovery tool vendors to consider

The vendors below sell cryptographic discovery or inventory capability today. The list is alphabetical, not exhaustive, and implies no ranking. Each description is taken from the vendor’s own product page as of November 2026. Some products are certificate lifecycle platforms with discovery features rather than full cryptographic inventories, which is exactly what the checklist above will show.

| Vendor | Product | What its own page says it discovers |
| --- | --- | --- |
| AppViewX | [Quantum Trust Hub](https://www.appviewx.com/products/quantum-trust-hub/) | Cryptography “across certificates, TLS configurations, cipher suites, protocols, and code libraries”, partly from existing Qualys and Tenable scans; source code scanning requires an agent. |
| DigiCert | [Trust Lifecycle Manager](https://www.digicert.com/trust-lifecycle-manager) | Certificates imported from any CA or trust store, and scans of “networks, clouds, and endpoints for unmanaged certificates”. |
| IBM | [IBM Quantum Safe](https://www.ibm.com/products/quantum-safe) | Cryptography “across IT operations and code repositories, including relevant interfaces and dependencies”. |
| Keyfactor | [AgileSec](https://www.keyfactor.com/products/cryptographic-discovery-inventory/) | Certificates, SSH keys, tokens, algorithms, protocols and cryptographic libraries across systems, endpoints, networks, cloud environments and code repositories. |
| QIZ Security | [QIZ Cryptography Management Platform](https://qizsecurity.com/) | “Continuous discovery” of cryptography “embedded across the organization, in services, protocols, and applications”, from on-premises to cloud, mapped against policy through API integrations (“No agents, no probes”). |
| Quantum Sentinel | [Quantum Sentinel](https://quantumsentinel.uk/product/) | Agentless, multi-tenant platform covering TLS, SSH and IPsec endpoints, certificate authorities, HSMs and key managers, cloud key services, databases, file shares, source code, endpoints (via CrowdStrike Falcon), Kubernetes, and operational technology and IoT. |
| SandboxAQ | [AQtive Guard](https://www.aqtiveguard.com/) | Keys, certificates and cryptographic operations, correlating data “across code, runtime, and the network”. |
| Tychon | [TYCHON Quantum Command](https://tychon.io/products/tychon/pqc-management-module/) | Cryptography found through “endpoint-based discovery, application-centric analysis, and passive network traffic inspection”, including hybrid and pure ML-KEM. |

Treat this list as a starting point for a shortlist. The checklist above and the pilot below are how to choose among them.

### One example against the checklist

To show what a plain statement of scope looks like, here is how Quantum Sentinel describes itself on its own pages. It calls itself “an agentless PQC discovery tool and Enterprise CBOM platform”, and its platform datasheet describes it as multi-tenant, with “native tenant isolation with a portfolio view for groups and service providers”. Its [cryptographic coverage page](https://quantumsentinel.uk/product/cryptographic-coverage/) lists eleven estate domains and says hybrid key exchange such as X25519MLKEM768 is “reported by name”, with a native probe for servers that only accept post-quantum groups. For OT and IoT, the same page says devices are “assessed by the same engine under profiles built for fragile equipment: non-disruptive by default, with low-and-slow pacing available and workers deployed inside the segment”, that MQTT over TLS brokers are covered, and that “IoT and OT are first-class asset types”. Its product page also lists an optional passive sensor on a mirror port, which matters where active scanning is not permitted (see our guide to [passive PQC discovery in OT](https://sinevis.com/passive-pqc-discovery-ot/)). Its pages describe discovery as agentless, read-only and metadata only, with private keys never retrieved, a read-only sidecar as the one exception in Kubernetes, and installation in air-gapped environments. Its CBOM page states that the inventory exports as CycloneDX 1.7 JSON.

The platform is equally clear about its limits. Its [automated discovery page](https://quantumsentinel.uk/solutions/use-cases/automated-cryptographic-discovery/) says it covers three of CISA’s four discovery types and does not inspect operating system packages, language dependencies or SBOMs, and that it runs “scheduled scanning, not continuous real-time monitoring”. Its other pages state that repository scanning reads the current state of one branch, not Git history, excludes common dependency directories and leaves credentials removed from history to secret-scanning tools, that file shares are SMB 2.0.2 to 3.1.1 and NFS v3 only, that it records host certificates rather than chains, and that the HSM it names is Thales Luna Network HSM. Those are the kinds of answers to expect from every vendor on a shortlist.

## How to run a fair 30-day pilot

A pilot is only fair if every tool sees the same scope under the same conditions, and you know the right answers in advance.

1. **Days 1 to 5: build the answer key.** Choose one bounded scope, such as a business service and its supporting infrastructure. Seed it with known items: a legacy certificate, an endpoint offering X25519MLKEM768, an SSH server on OpenSSH 10, an ML-KEM-only test server, an RSA key in an HSM test partition, keys in a cloud KMS test account, an encrypted archive with a misleading file name, and a repository with known algorithm calls. Write the list down before any tool runs.
2. **Days 1 to 5, in parallel: agree access.** Issue the same read-only credentials to each vendor, record the permissions each connector needed, and get change approval for any scanning windows.
3. **Days 6 to 15: run discovery.** Run each tool across the full scope, then run it again a week later without changes. The second run tests deduplication and drift handling.
4. **Days 16 to 22: score against the answer key.** Record what each tool found, missed and misreported, and verify a sample of findings outside the seeded items by hand.
5. **Days 23 to 27: test the output.** Export the CBOM, validate it against the CycloneDX schema, import a sample into your CMDB and raise a ticket from a finding.
6. **Days 28 to 30: decide.** Score each tool on what you saw, not what the vendor said. Note any gaps that will need another tool or manual work, as CISA’s strategy anticipates.

Keep the evidence: scope, answer key, raw exports and scoring sheet. It justifies the purchase and gives you a baseline for measuring the tool you choose.

## How Sinevis Can Help

Sinevis helps organisations write the requirements, build the answer key and run a side-by-side pilot of PQC discovery tools, as part of our [post-quantum cyber readiness](https://sinevis.com/cybersecurity-services/post-quantum-cyber-readiness/) service. For automated discovery and inventory we work with Quantum Sentinel, and we bring the consulting work no tool does: mapping findings to business services, data lifetime analysis and a migration plan your board can act on.

For a quick external baseline, try our free [PQC compliance check](https://sinevis.com/pqc-compliance-check/). To scope a discovery pilot, [contact the Sinevis team](https://sinevis.com/sinevis-contact/).

## Sources

- NIST, [NIST IR 8547, Transition to Post-Quantum Cryptography Standards (final)](https://csrc.nist.gov/pubs/ir/8547/final) (October 2026)
- UK National Cyber Security Centre, [Timelines for migration to post-quantum cryptography](https://www.ncsc.gov.uk/guidance/pqc-migration-timelines) (20 March 2025)
- Office of Management and Budget, [M-26-15, Execution of the Migration to Post-Quantum Cryptography](https://www.whitehouse.gov/wp-content/uploads/2026/06/M-26-15-Execution-of-the-Migration-to-Post-Quantum-Cryptography.pdf) (24 June 2026)
- CISA, [Strategy for Migrating to Automated Post-Quantum Cryptography Discovery and Inventory Tools](https://www.cisa.gov/resources-tools/resources/strategy-migrating-automated-post-quantum-cryptography-discovery-and-inventory-tools) (15 August 2024)
- IETF, [RFC 9846, The Transport Layer Security (TLS) Protocol Version 1.3](https://www.rfc-editor.org/rfc/rfc9846) (July 2026)
- IETF, [RFC 10024, Post-Quantum Traditional (PQ/T) Hybrid Key Agreement Mechanisms for TLS 1.3](https://www.rfc-editor.org/rfc/rfc10024) (August 2026)
- IANA, [Transport Layer Security (TLS) Parameters, TLS Supported Groups registry](https://www.iana.org/assignments/tls-parameters/tls-parameters.xhtml) (accessed 17 November 2026)
- OpenSSH, [OpenSSH release notes, OpenSSH 10.0](https://www.openssh.com/releasenotes.html) (9 April 2025)
- OWASP CycloneDX, [CycloneDX v1.7 released](https://cyclonedx.org/news/cyclonedx-v1.7-released) (21 October 2025)
- Ecma International, [ECMA-424, CycloneDX Bill of materials specification, 2nd edition](https://ecma-international.org/publications-and-standards/standards/ecma-424/) (December 2025)
- NIST, [SP 800-82 Rev. 3, Guide to Operational Technology (OT) Security](https://csrc.nist.gov/pubs/sp/800/82/r3/final) (September 2023)
- NIST, [SP 800-82 Rev. 4 (initial public draft)](https://csrc.nist.gov/pubs/sp/800/82/r4/ipd) (21 September 2026)
