This file gives AI coding assistants the context they need to work effectively in this repository.
deadline-cloud-samples is a public collection of samples for
AWS Deadline Cloud. It is not a single buildable package;
each sample is self-contained. The repository does have one top-level, standard-library-only static
validation command:
python3 scripts/validate_repository.pyRun it after every repository change, along with tests or validation owned by the sample you edit. External Markdown links use a separate network-dependent command:
python3 scripts/check_external_links.pyUse --no-ignore to audit the narrowly documented domain ignore list before changing it. Real
broken links must be fixed rather than ignored.
Start with the task paths and repository map in README.md, then use the nearest area or
collection README. Its index table is complete for the scope it declares. Search tracked paths or
contents directly when you need implementation or support files that are intentionally excluded from
user-selectable sample tables.
deadline-cloud-samples/
├── cloudformation/ CloudFormation farm and infrastructure templates
├── terraform/ Terraform farm and infrastructure templates
├── cdk/ AWS CDK farm apps (TypeScript)
├── job_bundles/ OpenJD job bundles (template.yaml plus assets)
├── conda_recipes/ DCC and application Conda recipes
├── containers/ Worker-compatible and application containers
├── queue_environments/ Session software environments (Conda, Rez, and pip)
├── host_configuration_scripts/ Privileged service-managed fleet setup scripts
├── submission_hooks/ Pre-submission Deadline Cloud CLI hooks
├── utility_scripts/ Standalone workflow helpers
├── skills/ Task-specific guides for coding agents
├── docs/ Contributor guidance and documentation starting points
└── scripts/ Standard-library repository validation
Read the relevant sample README.md before modifying its files. Use
docs/SAMPLE_README_TEMPLATE.md as an adaptable starting point when
adding a nontrivial sample.
Before implementing a sample, inspect skills/ for a matching SKILL.md guide.
- Use inclusive language: prefer
primary/replicaandallowlist/denylist. - Use
pip install ...in cross-platform Python instructions; avoidpip3unless the sample is explicitly Linux/macOS-only. - Job bundles live under
job_bundles/<name>/with atemplate.yaml, optionalparameter_values.yaml, and aREADME.mdfor nontrivial samples. - Conda recipes live under
conda_recipes/<package>-<version>/with arecipe/directory anddeadline-cloud.yaml. - For OpenJD templates, run
openjd checkandopenjd run --tasks <one>to verify a representative task locally before submitting the full parameter range when possible. - Add, rename, move, or delete a sample in the nearest category table. Change root navigation only when a recommended path changes.
- Keep indexes in Markdown. Do not add catalog or metadata-generation machinery.
- Do not add third-party runtime dependencies to repository validation.
- Run
python3 scripts/validate_repository.pysuccessfully (unit tests and static local-link checks). - Run
python3 scripts/check_external_links.pysuccessfully when Markdown links change. - Run the affected sample's own relevant tests or static checks.
- Update the sample README when behavior, prerequisites, parameters, outputs, or risks change.
- Update the nearest category table when a sample is added, renamed, moved, or deleted.
- Update root task paths only when a recommended starting point changes.
- Use a conventional commit title.
- Sign off every commit under the Developer Certificate of Origin.
- Check changed content for inclusive language.
See CONTRIBUTING.md for the full contribution and licensing workflow.