| name | develop-release-integration |
| description | Build, test, and maintain this container-based integration plugin for Digital.ai Release (Python SDK). Use for project setup, defining task types and server connections in resources/type-definitions.yaml, implementing tasks in src/, writing pytest tests, and the build/deploy workflow. Routes to the detailed guide; does not duplicate it. |
| license | MIT |
| metadata | {"audience":"developers","workflow":"digital-ai-release"} |
Develop a Release container plugin
A portable, tool-neutral skill for working in this template. It is concise on purpose: the
authoritative content lives in the repo docs, and this file routes you to the right one so
nothing is duplicated or drifts.
- AGENTS.md — conventions and guardrails. Read first.
- PLUGIN_DEVELOPMENT.md — the detailed guide (architecture, type
definitions, task patterns, examples, dev environment, troubleshooting, Kubernetes).
- README.md — setup, build, and install commands.
What this covers
Setting up the project · defining task types and server connections in
resources/type-definitions.yaml · implementing tasks in src/ · writing pytest tests ·
building the plugin zip + Docker image and installing into Release.
The one rule you must not break
A task's type maps to its Python class by the name after the dot
(containerExamples.Reverse → class Reverse). The class name must match exactly and be
unique across src/. Without scriptLocation, resolution is by class name; with
scriptLocation, the path must point to the exact file under src/.
→ details
Route the request
Always
- Keep
resources/type-definitions.yaml and src/ classes in lockstep (the naming contract).
- Add a runtime dependency to both
requirements.txt (image) and pyproject.toml (dev).
- Tests are pytest in
tests/unit (fast) and tests/integration (networked; Release-backed
tests auto-skip, third-party service tests require internet access).
Run uv run pytest before building (on Windows, add --python .venv/Scripts/python.exe if
uv warns about VIRTUAL_ENV).