kanikorevanced/executor repository overview

⁠kaniko-revanced: Modern, Fast & Secure Image Builder in Userspace

kaniko-revanced is an actively maintained, high-performance, and drop-in compatible continuation of the original Google Kaniko (GoogleContainerTools/kaniko). It executes container image builds from a Dockerfile inside a container or Kubernetes cluster without requiring a Docker daemon or privileged security contexts.

⁠🏷️ Image Tags & Use Cases

All images are multi-architecture (linux/amd64, linux/arm64, linux/s390x, linux/ppc64le).

⁠🤖 AI-Powered Build Engine (New in v0.2.0)

kaniko-revanced includes native AI intelligence directly inside the builder. It connects to any OpenAI-compatible endpoint (Local Ollama, Google Gemini, OpenAI, Anthropic, Groq, DeepSeek, OpenRouter, or enterprise gateways).

⁠Key AI Features:

• 🔧 Autonomous Auto-Healing (--llm-auto-heal): Catches failing RUN/COPY steps in real-time, prompts the LLM for a validated fix, prints a colorized unified diff, and automatically retries the build up to --llm-max-retries=N.
• 📉 Deep Diagnostics & Optimization (--llm-diagnose): Pinpoints the exact root-cause of build failures, suggests image size reductions (multi-stage builds, cache purging), and audits security misconfigurations.
• ⚡ Apply Optimizations Directly (--llm-diagnose-apply): Translates optimization advice into a production-ready patched Dockerfile.
• 🛡️ Pre-Flight Health Probes & Fail-Open Safety: Performs non-blocking reachability checks. If an AI endpoint is missing or offline, Kaniko issues a warning and proceeds with a standard deterministic build without failing CI.
• 🔒 Zero-Leakage Privacy Sanitization: Automatically redacts passwords, tokens, AWS credentials, and secrets before prompts leave your environment.
• 📑 GitLab CI Log Protection: Automatically wraps diagnostic reports in native collapsible sections (\e[0Ksection_start...) to prevent exceeding GitLab's 4MB job log limit.

⁠🛠️ Quickstart Examples

⁠1. GitLab CI (.gitlab-ci.yml)

build:
  stage: build
  image:
    name: kanikorevanced/executor:debug
    entrypoint: [""]
  script:
    - mkdir -p /kaniko/.docker
    - echo "{\"auths\":{\"$CI_REGISTRY\":{\"username\":\"$CI_REGISTRY_USER\",\"password\":\"$CI_REGISTRY_PASSWORD\"}}}" > /kaniko/.docker/config.json
    - /kaniko/executor
        --context "${CI_PROJECT_DIR}"
        --dockerfile "${CI_PROJECT_DIR}/Dockerfile"
        --destination "${CI_REGISTRY_IMAGE}:${CI_COMMIT_TAG:-latest}"

⁠2. GitLab CI with AI Auto-Healing

build-with-ai:
  stage: build
  image:
    name: kanikorevanced/executor:ai-debug
    entrypoint: [""]
  variables:
    KANIKO_LLM_API: "https://api.groq.com/openai/v1"
    KANIKO_LLM_KEY: "$GROQ_API_KEY"
    KANIKO_LLM_MODEL: "llama-3.3-70b-versatile"
    KANIKO_LLM_AUTO_HEAL: "true"
    KANIKO_LLM_DIAGNOSE: "true"
  script:
    - mkdir -p /kaniko/.docker
    - echo "{\"auths\":{\"$CI_REGISTRY\":{\"username\":\"$CI_REGISTRY_USER\",\"password\":\"$CI_REGISTRY_PASSWORD\"}}}" > /kaniko/.docker/config.json
    - /kaniko/executor
        --context "${CI_PROJECT_DIR}"
        --dockerfile "${CI_PROJECT_DIR}/Dockerfile"
        --destination "${CI_REGISTRY_IMAGE}:latest"

⁠3. Kubernetes Pod

apiVersion: v1
kind: Pod
metadata:
  name: kaniko-builder
spec:
  containers:
  - name: kaniko
    image: kanikorevanced/executor:latest
    args:
      - "--context=git://github.com/my-org/my-repo.git#refs/heads/main"
      - "--destination=myregistry.azurecr.io/my-app:latest"
    volumeMounts:
      - name: docker-config
        mountPath: /kaniko/.docker
  restartPolicy: Never
  volumes:
    - name: docker-config
      secret:
        secretName: regcred
        items:
          - key: .dockerconfigjson
            path: config.json

⁠4. Local Docker CLI with Local LLM (Ollama)

docker run --rm -it --net=host \
  -v $(pwd):/workspace \
  kanikorevanced/executor:ai \
  --context /workspace \
  --dockerfile /workspace/Dockerfile \
  --no-push \
  --llm-auto-heal \
  --llm-api="http://localhost:11434/v1" \
  --llm-model="qwen2.5-coder:7b"

⁠⚡ Performance Advantages

In benchmark comparisons against upstream Google Kaniko on complex multi-stage workloads:

• Average Build Speed: 52.4% faster (2.1x speedup) through reusable sync.Pool memory allocation and optimized tar serialization.
• CVE Resolution: Proactively patched against all known CVEs in upstream toolchains.
• 100% Drop-In Compatible: Replaces gcr.io/kaniko-project/executor without modifying arguments.

⁠📄 License & Source Code

• GitHub Repository: https://github.com/raj-sh-git/kaniko-revanced⁠
• License: Apache License 2.0