nf-core AGENTS check example

User: Read this file: https://raw.githubusercontent.com/nf-core/agents/refs/heads/add-agents-md/AGENTS.md. Confirm that you have the file and that you can read it.

GitHub Copilot: I’ll fetch that AGENTS.md URL now and confirm I can read its contents.

Fetched https://raw.githubusercontent.com/nf-core/agents/refs/heads/add-agents-md/AGENTS.md

Confirmed. I successfully fetched that file and can read it.

I can access:
- https://raw.githubusercontent.com/nf-core/agents/refs/heads/add-agents-md/AGENTS.md

I also parsed substantive content from it (sections like “Nextflow pitfalls”, “nf-core template structure”, “nf-test and testing”, and “git and branch policy”), so read access is working.

User: Great. Quiz time. I ask, you answer, no coding needed. Question 1: Under what circumstances can you write code to the TEMPLATE branch?

GitHub Copilot: Never. You cannot write code to the TEMPLATE branch under any circumstances; it is explicitly managed by nf-core tools and must not be written to directly.

User: How many approvals are required to merge a PR to the main branch?

GitHub Copilot: Two approvals are required to merge a PR to the `main` branch.

User: Which keyword do you use to define a variable in Nextflow?

GitHub Copilot: Use `def` to define a variable in Nextflow.

User: I need to find the output file paths for a module called STRELKA in a pipeline with legacy outputs. Where should I look for it?

GitHub Copilot: For legacy outputs, check the places where `publishDir` is defined for that process.

Start here:
1. The STRELKA module process definition itself (its process block often contains `publishDir`).
2. The central process configuration overrides for modules (where `withName` selectors can set or override `publishDir` for STRELKA).

In short: find the `publishDir` attached to STRELKA, because in legacy-output pipelines that is what determines the final output paths.

User: Suppose you want to add a new tool to the pipeline. Which steps would you take and which files could be affected?

GitHub Copilot: 1. Decide integration strategy first:
- Reuse an existing nf-core module if possible.
- If no suitable reusable module exists, create a local module with nf-core tools.

2. Add the module or subworkflow with nf-core tooling:
- Install/create with nf-core commands rather than hand-crafting boilerplate.
- If you must adjust an nf-core module implementation, patch it afterwards with nf-core modules patch.

3. Wire the tool into pipeline execution:
- Connect inputs/outputs in the main workflow graph in nf-amplivar.nf and, if needed, orchestration in main.nf.
- Keep output style consistent with the pipeline’s existing output system (legacy publishDir vs workflow outputs).

4. Configure runtime options and resources:
- Add module-specific settings (for example ext.args/ext.prefix, resources) in modules.config.
- Update global/default params if needed in nextflow.config and profile-specific configs such as test.config or test_full.config.

5. Expose and validate new parameters:
- Extend the schema in nextflow_schema.json.
- If input format changes are needed, update schema_input.json.

6. Add or update tests:
- Add/update pipeline-level tests in default.nf.test and test config in nextflow.config.
- Add/update module tests in the corresponding module test area under the modules directory.

7. Update documentation:
- Document usage/params/outputs in README.md.

Most commonly affected files in this repository are:
- nf-amplivar.nf
- main.nf
- modules.config
- nextflow.config
- nextflow_schema.json
- default.nf.test
- nextflow.config
- README.md