DETERMINISTIC SECURITY AUDIT

KCode
Bug-scan your codebase in 10k tokens, not 300k.

A deterministic security scanner for C, Rust, Go, Python, and 20+ other languages. 414 curated patterns — hand-written and regression-tested. Ran it on NASA's nasa/IDF flight-software repo and found real bugs (PR #107 submitted, pending review). The LLM only verifies; the scanner does the work. Runs 100% on your hardware. SARIF output for CI.

BenchmarksGitHub

Head-to-head with Semgrep, Semgrep Pro, CodeQL on OWASP Benchmark v1.2 →

KCODE
$ kcode audit .
KCode Audit Engine
Project: /home/dev/myapp
Output: AUDIT_REPORT.md
[Verifier] mark7 → localhost:10091 (local · 31B params)
scanning: 1,247 files · 414 patterns
verifying: 37 candidates
[1/37] py-002-shell-injection — workers.py:41
[2/37] java-023-sql-injection-var-flow — UserController.java:52
[3/37] go-005-cmdi-command-flow — handlers.go:88
[4/37] cpp-008-strcpy-unbounded — net/parser.cpp:120
Report written: AUDIT_REPORT.md · AUDIT_REPORT.json
Files scanned: 1247 / 1247
Candidates found: 37
Confirmed findings: 28
High: 4 Medium: 12 Low: 12
False positives: 9
Duration: 42.3s

From code to pull request in three commands

The audit pipeline is a first-class primitive, not a chat instruction you have to negotiate with the model.

/scan

Find candidates

Pattern library runs deterministically across your tree. Optional LLM verifier downgrades false positives. Produces AUDIT_REPORT.md + SARIF.

/fix

Apply patches

Each pattern ships with a fix template. Size guards, bounded copies, RAII wrappers applied automatically. Diff-previewed before write.

/pr

Open the PR

Branch + commit + LLM-generated PR description grounded in the finding evidence. Auto-fork and submit when you don't own the repo.

Built to find real bugs, not hallucinate them

LLM-first tools send your whole codebase to a cloud model and hope it notices the bug. KCode flips it: a deterministic scanner pre-filters to ~10k tokens of actual candidates, then a local LLM verifies. Faster, cheaper, provable.

414 Curated Patterns

Not LLM-generated. Every pattern is hand-written, ships with positive + negative fixtures, and survives a CI regression harness on every release.

10k Tokens per Audit

The deterministic scanner does the heavy lifting. The LLM only verifies — seeing ~10k tokens of pre-filtered candidates instead of 300k+ of raw source.

20+ Languages

C, C++, Rust, Go, Python, Java, JavaScript, TypeScript, PHP, Ruby, Swift, Kotlin, C#, Scala, Haskell, Zig, Dart, Lua, SQL — plus framework packs for Flask, Rails, React.

SARIF + GitHub Action

Output is SARIF v2.1.0 — drop-in compatible with GitHub Code Scanning. Findings appear as inline PR comments. Ships as a one-line GitHub Action.

Real bugs in NASA OSS

Found on nasa/IDF: pointer arithmetic, unreachable code, resource leaks. PR #107 has the diffs. Not a synthetic benchmark — production-grade flight-software code.

100% Local & Private

Audit runs on your GPU. Source never leaves your machine. Cloud LLM verification is optional, not required. No telemetry, no upload.

Local verifier: mnemo:mark6

KCode's scanner produces the candidates; a small local model verifies them. Gemma 4 31B abliterated runs on consumer hardware — your source never leaves the machine. Auto-selected per your GPU.

ModelSizeMin VRAMBest for
mnemo:mark6-pico2.6 GB3 GBCompact 4B abliterated — 4GB GPUs or CPU
mnemo:mark6-nano6 GB12 GBFast dense 8B — 12GB GPUs
mnemo:mark6-mini13.5 GB16 GBGemma 4 dense 14B — 16GB GPUs
mnemo:mark6-mid18.6 GB24 GBGemma 4 31B Q4_K_M — 24GB GPUs
mnemo:mark6-max33 GB36 GBGemma 4 31B Q8_0 — best quality, 32GB+ GPUs
mnemo:mark6-31b33 GB36 GBFull Gemma 4 31B abliterated — flagship model

GGUF format — works on NVIDIA (CUDA), AMD (ROCm), Apple Silicon (Metal), and CPU

Download KCode v2.10.466

Single binary, no dependencies. The setup wizard handles everything automatically.

Free & open-source (Apache 2.0). The setup wizard handles everything automatically.

Pay only for the hard bugs

The scanner is free and open source. Pro adds cloud verification for tricky cases and the /fix + /pr automation. No cloud lock-in, no telemetry, no code upload.

KCode Free

$0

Full scanner, unlimited local audits, community support

  • 414 patterns across 20+ languages
  • Unlimited local audits
  • SARIF + Markdown output
  • Local LLM verifier (mnemo:mark6)
  • Apache 2.0 open source
MOST POPULAR

KCode Pro

$19/mo

Cloud verification, /fix + /pr, priority support

  • Everything in Free
  • Cloud LLM verifier (higher accuracy on complex bugs)
  • /fix — deterministic patches
  • /pr — auto-branch + commit + description
  • GitHub Action for CI
  • Priority support

KCode Solo

$99 one-time

Everything in Pro, permanent license, no subscription

  • Everything in Pro
  • Permanent license
  • No subscription ever
  • Lifetime updates
  • Priority support

Volume licensing: contact sales

Your bugs. Your audit. Your machine.

414 patterns, 20+ languages, 10k tokens per run, SARIF out of the box. Source never leaves your hardware. Download KCode and audit your repo in a single command.

View on GitHub