By licongxu
Research paper-writing, plotting validation, coding, and FLAMINGO data skills for scientific computing workflows.
Reviews git diffs against plan/requirements using the code-reviewer skill. Use when the user asks for a code review, before merge, or after a major feature.
FLAMINGO simulation expert. Use when the user wants to locate, select, read, or analyse FLAMINGO data products (simulations, snapshots, SOAP halo catalogues, lightcones, HEALPix maps, power spectra) — including remote streaming access via hdfstream / swiftsimio.
Audits and tidies repository layout using the repo-hygiene skill. Use when the user asks to organize, consolidate, or clean up a repo or folder.
Review code changes against requirements and quality standards before merge. Use when dispatching a review subagent, after major features, before merging, or when the user asks for a code review. Requires reading the actual diff.
Test-driven workflow: write failing tests first, verify red/green, cover edge cases, and keep the full suite green before merge. Use when implementing features, fixing bugs, or refactoring — before and alongside production code.
Guidelines for implementing code changes: think first, keep diffs minimal and simple, surface assumptions, and verify with tests including edge cases. Use when writing new code, fixing bugs, or refactoring — before and during implementation.
Audit and tidy a repo or working folder: fix layout drift, consolidate misplaced files, remove stale artifacts, align docs with disk. Use when organizing any project, scratch folder, or new repo — before release or after a refactor.
Deep-read an academic paper and produce a durable, critical research note — not a loose summary. Use when given a paper title, DOI, arXiv ID, publisher/PDF/GitHub URL, or local PDF, or when asked to read closely, summarize, extract contributions, analyze experiments, critique, compare to related work, reproduce results, or maintain a paper-reading repository index. Works across fields (astro/physics, ML, bio, security, …); examples lean astro-ph but the workflow is general.
Modifies files
Hook triggers on file write and edit operations
Uses power tools
Uses Bash, Write, or Edit tools
Own this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimOwn this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimBased on adoption, maintenance, documentation, and repository signals. Not a security audit or endorsement.
Portable skills for Claude Code, Cursor, Ralph loops, and other agent workflows.
This repository is also a Claude Code plugin — skills live under skills/; manifest
under .claude-plugin/ (tracked in git).
Design: principles and checks, not copy-paste recipes. Every SKILL.md works
standalone in any project — symlink, copy, or install via plugin. Project-specific
commands, units, and science belong in that repo's CLAUDE.md.
cosmo-agent-skills/
├── hooks/ # plugin Stop + PreToolUse hooks
├── commands/ # /loop-on, /loop-off (plugin slash commands)
├── scripts/ # loop toggle helpers
├── loop-prompt.md # default continuation checklist
├── .claude-plugin/
│ ├── plugin.json # plugin manifest
│ └── marketplace.json # marketplace catalog
├── agents/ # bundled subagents (code-reviewer, repo-organizer)
├── examples/ # settings snippets
├── ref_papers/ # local PDFs only (gitignored)
└── skills/
├── hydrosim/ # simulation data (FLAMINGO: 6 skills)
├── plotting/ # figure export + validation (6 skills)
├── paper-reading/ # deep-read papers → critical notes (1 skill)
├── paper-writing/ # manuscript workflow (3 skills)
└── coding/ # software engineering (4 skills)
Add new categories under skills/ with a category README.md.
Optional autonomous iteration: bundled in this plugin. A Stop hook re-sends the loop prompt whenever Claude tries to end a turn, until you run /loop-off.
/plugin install cosmo-agent-skills@cosmo-agent-skills)./loop-on, then your first task. /loop-off when done.Prompt source: plugin default loop-prompt.md, or override with .claude/loop-prompt.md in your project (created by /loop-on with arguments).
Requires bash, jq, and realpath on the host. Optional: merge examples/claude-settings.nonstop-loop.json if long prompts are truncated.
Safety: while /loop-on is active, Claude cannot exit until /loop-off. protect-write-scope limits Edit/Write to CLAUDE_PROJECT_DIR.
Per Create plugins:
# Dev: load this repo as a plugin
claude --plugin-dir .
# Persistent: register marketplace + install
/plugin marketplace add /path/to/cosmo-agent-skills
/plugin install cosmo-agent-skills@cosmo-agent-skills
Skills invoke as /cosmo-agent-skills:<skill-name> (e.g. /cosmo-agent-skills:code-testing).
Validate: claude plugin validate .
Bump version in .claude-plugin/plugin.json and marketplace.json when releasing.
For repos that vendor skills under .claude/skills/ (e.g. autoflamingo at /scratch/scratch-lxu/flamingo_data_analysis/autoflamingo):
/path/to/cosmo-agent-skills/scripts/sync-skills-to-project.sh /path/to/your/project
Re-run after pulling updates to this repository.
Cursor / Claude Code (standalone skills) — symlink or copy individual skills:
git clone https://github.com/licongxu/cosmo-agent-skills.git
cd cosmo-agent-skills
# Cursor: one skill → ~/.cursor/skills/
for d in skills/plotting/*/ skills/paper-writing/*/ skills/coding/*/; do
ln -sf "$(pwd)/$d" ~/.cursor/skills/"$(basename "$d")"
done
Skills trigger by name/description in any repo. No dependency on this repository's layout.
Claude Code plugin — see Claude Code plugin above for
/plugin install and namespaced skills (/cosmo-agent-skills:code-testing).
FLAMINGO data release — run inventory and data products. Facts verified against the portal and Schaye et al. (2023) / Helly et al. (2026) release paper.
| Skill | Use when |
|---|---|
| flamingo-simulations | Choosing a run, paths, naming, variations |
| flamingo-power-spectra | ASCII P(k), emulator, units |
| flamingo-snapshots | Particle HDF5, partial snapshots |
| flamingo-halo-catalogues | SOAP / HBT-HERONS catalogues |
| flamingo-lightcones | HEALPix maps, particle and halo lightcones |
| flamingo-hdfstream | Remote streaming access to FLAMINGO HDF5 on COSMA (default) |
npx claudepluginhub licongxu/cosmo-agent-skills --plugin cosmo-agent-skillsComprehensive skill pack with 66 specialized skills for full-stack developers: 12 language experts (Python, TypeScript, Go, Rust, C++, Swift, Kotlin, C#, PHP, Java, SQL, JavaScript), 10 backend frameworks, 6 frontend/mobile, plus infrastructure, DevOps, security, and testing. Features progressive disclosure architecture for 50% faster loading.
Harness-native ECC operator layer - 67 agents, 271 skills, 92 legacy command shims, reusable hooks, rules, selective install profiles, and production-ready workflows for Claude Code, Codex, OpenCode, Cursor, and related agent harnesses
Access thousands of AI prompts and skills directly in your AI coding assistant. Search prompts, discover skills, save your own, and improve prompts with AI.
Upstash Context7 MCP server for up-to-date documentation lookup. Pull version-specific documentation and code examples directly from source repositories into your LLM context.
Complete developer toolkit for Claude Code
Develop, test, build, and deploy Godot 4.x games with Claude Code. Includes GdUnit4 testing, web/desktop exports, CI/CD pipelines, and deployment to Vercel/GitHub Pages/itch.io.