Appearance

Quickstart ​

Build the Veloce agent, run it locally, and make your first ML-KEM and ML-DSA calls.

Building and testing is currently supported on Linux x86-64.

Build and test everything ​

bash scripts/run_gates.sh

This builds the wolfCrypt FIPS module from the licensed bundle (vendor/wolfssl), builds the PQC provider, the agent, qSearch, and the CLI, generates ~/.veloce/agent.json, and runs the release-gate battery. Expect ALL GATES GREEN.

To force a clean rebuild:

FORCE_REBUILD=1 bash scripts/run_gates.sh

Run the agent ​

python3 scripts/gen_config.py
build/bin/veloce-agent --config ~/.veloce/agent.json &

Use the Python SDK ​

The SDK is pure Python — no cryptographic code runs in-process. Every call executes inside the local agent over an authenticated IPC channel.

python
import sys; sys.path.insert(0, "python")   # or: pip install build/dist/*.whl
import veloce

veloce.initialize()

kp = veloce.mlkem_generate_keypair()               # ML-KEM-768
ct, ss = veloce.mlkem_encapsulate(kp.public_key)
assert veloce.mlkem_decapsulate(kp.private_key_handle, ct) == ss

sk = veloce.mldsa_generate_keypair()               # ML-DSA-65
sig = veloce.mldsa_sign(sk.private_key_handle, b"msg")
assert veloce.mldsa_verify(sk.public_key, b"msg", sig)

veloce.export_cbom(format="cyclonedx", path="cbom.cdx.json")

Set VELOCE_SOCKET to point the SDK at a non-default agent socket path.

Use the CLI ​

build/bin/veloce status
build/bin/veloce validation
build/bin/veloce self-test
build/bin/veloce cbom cyclonedx
build/bin/veloce scan /path/to/codebase

Run qSearch directly ​

build/bin/qsearch scan /path/to/codebase --out qsearch-out
build/bin/qsearch system --out host-inventory

See Cryptographic discovery for what each mode inventories and the report formats produced.

Next steps ​