| name | nanocortex |
| description | Execute Oxford Nanopore and RNA workflows in Cursor with NanoCortex's Singularity image, including Dorado, Modkit, transcript/splicing analysis, RNA-FM structure prediction, and Remora signal plotting. Use when the user mentions NanoCortex, nanopore reads, BAM/BED modifications, GTF isoforms, RNA FASTA, POD5 signal plots, Singularity workflows, or asks to run @nanocortex. |
| disable-model-invocation | true |
NanoCortex for Cursor
Install this directory at ~/.cursor/skills/nanocortex (personal) or .cursor/skills/nanocortex (project). Invoke explicitly with @nanocortex.
Open the NanoCortex repository as the workspace when possible. Default container: singularity/bot.sif. Do not install alternative software when the container is available.
Read references/environment.md only when the direct Singularity command is unavailable.
Cursor execution model
- Use the integrated terminal for shell commands. Keep commands copy-pasteable and minimal.
- Prefer workspace-relative paths when the NanoCortex repo is open; otherwise use absolute paths.
- Read only the selected reference, script
--help, and named input files. Do not recursively search the repo.
- On failure, make one targeted correction from stderr, then retry once before escalating.
Fast path
- Identify the task from the request and input suffix.
- Read exactly one reference:
- Prefer bundled scripts. Run
python3 scripts/<script>.py --help from this skill directory.
- Verify the selected container tool once with
--help or --version, then execute.
- Validate outputs with the bundled validator or the checks in the selected reference.
- Summarize method, parameters, main results, output paths, and limitations.
For Dorado, Modkit, or RNA-FM option lookup, use scripts/query_parameters.py with the bundled JSON specifications.