By quantumbfs
Generate and submit sbatch scripts to run GPU compute jobs on Slurm clusters, selecting GPU types like A100_40G, V100, or A800, configuring partitions, nodes, Python paths, and cluster-specific rules for efficient job execution and log management.
npx claudepluginhub quantumbfs/claude-code-skills --plugin submit-slurm-jobA small marketplace of Claude Code plugins, each shipping one skill.
In Claude Code, add this repo as a marketplace, then install whichever plugin(s) you want:
/plugin marketplace add QuantumBFS/claude-code-skills
/plugin install download-papers@quantum-bfs
/plugin install submit-slurm-job@quantum-bfs
/plugin install verify-references@quantum-bfs
The @quantum-bfs suffix matches the name field in .claude-plugin/marketplace.json.
To update later:
/plugin marketplace update quantum-bfs
download-papersDownload an academic paper as a PDF given a URL, DOI, paper title, or arXiv ID. Resolves in order: open web (author homepage, repositories) → Sci-Hub mirrors → arXiv.
Triggers on phrases like "download paper", "get me the PDF", "fetch this DOI". No configuration required.
submit-slurm-jobGenerate and submit sbatch scripts for GPU compute jobs on a Slurm cluster. Handles partition selection, explicit GPU model in --gres, full Python path, and log file management.
Required configuration — add to your project's CLAUDE.md:
## Slurm Configuration
- PYTHON_PATH: `/path/to/miniconda3/envs/myenv/bin/python3`
- PROJECT_DIR: `/home/your_user/private/homefile` # scripts/logs must live here
- PARTITION: `home` # compute partition
Invoke with /submit-slurm-job (or just describe the job — Claude will pick up the skill).
verify-referencesVerify a BibTeX bibliography against CrossRef metadata. The skill:
.tex files to flag uncited entries..bib.Useful before submission to catch hallucinated or stale BibTeX. Invoke by asking to "verify references" / "check the bib file". Requires Python 3 and an internet connection (for the CrossRef API).
.
├── .claude-plugin/
│ └── marketplace.json # marketplace manifest, lists all plugins
└── plugins/
├── download-papers/
│ ├── .claude-plugin/plugin.json
│ └── skills/download-papers/SKILL.md
├── submit-slurm-job/
│ ├── .claude-plugin/plugin.json
│ └── skills/submit-slurm-job/SKILL.md
└── verify-references/
├── .claude-plugin/plugin.json
└── skills/verify-references/
├── SKILL.md
├── check_unused_refs.py
├── compare_refs.py
└── download_crossref.py
Helper scripts shipped alongside a SKILL.md are referenced from the skill instructions via ${CLAUDE_SKILL_DIR}, which Claude Code expands to the skill's install directory at invocation time.
Add a new plugin by:
plugins/<your-plugin>/.claude-plugin/plugin.json (use one of the existing files as a template).plugins/<your-plugin>/skills/<your-plugin>/SKILL.md with YAML frontmatter (name, description)..claude-plugin/marketplace.json.Keep skills well-scoped, parameterize any user-specific paths, and document required configuration in this README.
MIT
Download academic papers as PDFs given a URL, DOI, title, or arXiv ID. Tries open web → Sci-Hub → arXiv.
Research integrity plugin for Claude Code — paper auditing, citation verification, experiment analysis, and methodology-first skills for academic workflows.
MCP server for Slurm workload management and HPC job scheduling
Oh My Paper research harness: memory system, Codex delegation, and pipeline commands for academic research projects.
Annotated research paper collection management — retrieve, read, extract, cross-reference scientific papers
Share bugs, ideas, or general feedback.
Ultra-compressed communication mode. Cuts ~75% of tokens while keeping full technical accuracy by speaking like a caveman.
Own this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge.
Sign in to claimOwn this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge.
Sign in to claim