agentleFS
rrezartprebrezarrezartprebreza/spring-boot-skills/skills/spring-boot-3/container-native-deployment/SKILL.md

container-native-deployment

Skill344 starsChanged 3 months agoMITView on GitHub

Use when packaging a Spring Boot 3 application as an OCI image or GraalVM native executable. Covers buildpacks, layered images, JVM containers, AOT hints, probes, security, and verification.

The file

We split this skill into the rules it gives an agent

12 rules. About 0 are common, 11 are distinctive, and 1 is very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈11

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 1

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Prefer the Spring Boot build-image task with Cloud Native Buildpacks for standard services.

    line 14
  2. Distinctive:

    Keep dependency and application layers separate to maximize cache reuse.

    line 15
  3. Distinctive:

    Run as a non-root user on a read-only filesystem where the application permits it.

    line 17
  4. Distinctive:

    Keep startup probes for slow initialization instead of weakening liveness.

    line 25
  5. Distinctive:

    Write temporary files only under an explicit writable location.

    line 26
  6. Distinctive:

    Use Spring AOT and the supported GraalVM native build tools.

    line 30
  7. Very specific:

    Add RuntimeHints for reflection, resources, serialization, or proxies that analysis cannot infer.

    line 31
  8. Distinctive:

    Avoid runtime bean-shape changes and other closed-world violations.

    line 32
  9. Distinctive:

    Run native integration tests.

    line 33
  10. Distinctive:

    Use immutable image digests for promotion between environments.

    line 38
  11. Distinctive:

    Verify signatures against the expected signer identity and issuer or trusted key, and check that provenance refers to the intended source …

    line 42
  12. Distinctive:

    Test untrusted-signer and digest-mismatch rejection before enforcing a deployment policy.

    line 46

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Container and Native Deployment

Choose JVM or native from measured startup, memory, throughput, and build constraints.

Container image

  • Prefer the Spring Boot build-image task with Cloud Native Buildpacks for standard services.
  • Keep dependency and application layers separate to maximize cache reuse.
  • Pin builder and run-image families in reproducible release pipelines.
  • Run as a non-root user on a read-only filesystem where the application permits it.
  • Supply secrets at runtime; never bake credentials or environment files into an image.
  • Configure graceful shutdown and align platform termination grace with application timeouts.

Runtime behavior

  • Set memory and CPU limits, then verify JVM ergonomics inside those limits.
  • Expose dedicated readiness and liveness probes through Actuator.
  • Keep startup probes for slow initialization instead of weakening liveness.
  • Write temporary files only under an explicit writable location.

Native image

  • Use Spring AOT and the supported GraalVM native build tools.
  • Add RuntimeHints for reflection, resources, serialization, or proxies that analysis cannot infer.
  • Avoid runtime bean-shape changes and other closed-world violations.
  • Run native integration tests; a successful JVM test suite is not sufficient.

Supply chain

  • Generate an SBOM, scan the final image, and patch builder/run images regularly.
  • Use immutable image digests for promotion between environments.

Scan the actual final image by immutable digest, including OS packages and bundled Java dependencies; record the scanner database timestamp and findings that need manual triage. Verify signatures against the expected signer identity and issuer or trusted key, and check that provenance refers to the intended source commit and build workflow. A signature from an arbitrary identity is not a release policy. Retain SBOM/provenance evidence for the promoted digest; rebuilding the same tag produces a different artifact and requires new verification. Test untrusted-signer and digest-mismatch rejection before enforcing a deployment policy. For dependency/BOM remediation and CI trust-boundary review, use java-supply-chain-security if installed; this skill remains responsible for image construction and runtime behavior.

Examples

  • See examples/good-dockerfile and examples/bad-dockerfile.

The good example targets Boot 3.3+ and uses jarmode=tools extraction in a builder stage. For Boot 3.0-3.2, use the older -Djarmode=layertools ... extract command. A normal Maven package does not create target/dependencies/ or target/application/ directories by itself.

Official sources

Gotchas

  • Agent copies one fat jar into a mutable root container - use layers and a non-root runtime.
  • Agent bakes secrets into image layers - inject secrets only at runtime.
  • Agent adds broad reflection configuration to make native builds pass - register narrow runtime hints.
  • Agent uses liveness to test every dependency - use readiness for traffic-affecting dependencies.
  • Agent chooses native without measuring throughput and build cost - benchmark both deployment modes.
  • Agent copies nonexistent target/dependencies directories - extract the packaged jar in a builder stage.
  • Agent uses jarmode=tools on Boot 3.0-3.2 - use layertools until the project reaches Boot 3.3.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.