Stats
Links
Categories
Claude Code plugins for code quality, scientific computing, and Bayesian inference workflows.
npx claudepluginhub matthiaskloft/claude-skillsDomain-specific skills for BayesFlow 2.x extension packages. Provides patterns for simulators, adapters, Keras 3 backend-agnostic math, testing, SBC validation, GPU memory management, and packaging conventions.
Plan/Implement/Ship workflow for feature development. Provides skills for structured planning, worktree-based implementation, and automated review/PR/CI shipping.
A monorepo of Claude Code plugins for scientific computing, code review, and workflow automation.
Domain-specific skills for BayesFlow 2.x extension packages. Provides patterns and best practices for building BayesFlow extensions.
| Skill | Description |
|---|---|
| bayesflow-simulator | Data-generating processes with @allow_batch_size, RNG discipline, and prior/likelihood/meta functions |
| keras-ops | Backend-agnostic tensor math using keras.ops.* for loss functions, layers, and forward passes |
| bayesflow-adapter | Data preprocessing pipelines, three-slot schema, and standardization transforms |
| bayesflow-testing | Test patterns with conftest.py setup, shape testing, and mocking |
| bayesflow-validation | Simulation-based calibration (SBC), coverage metrics, and quality gates |
| bayesflow-memory | GPU memory management, OOM recovery, and gradient checkpointing |
| bayesflow-packaging | Package structure, __all__ exports, pyproject.toml, and CI configuration |
Skills trigger automatically based on task context (e.g., creating a simulator, debugging OOM errors, writing tests). You can also invoke them explicitly via the bayesflow-skills: namespace.
Independent code review via specialized subagents. Reviews uncommitted changes for bugs, security issues, silent errors, and logical inconsistencies. Specialized for R and Python codebases.
| Skill | Description |
|---|---|
| code-sentinel | Orchestrates independent review subagents that audit your diff with no prior context of the implementation |
Usage: Ask Claude to "review my changes", "check my code before committing", or "audit this diff". The skill delegates to subagents that have no knowledge of the implementing conversation, ensuring unbiased review.
A Plan/Implement/Ship workflow for structured feature development. Breaks large features into phases, implements them in worktrees, and ships each phase as a PR with CI monitoring.
| Skill | Description |
|---|---|
| plan | Create a structured design document before implementation — scope, decisions, and phased steps |
| implement | Build from a plan document phase by phase, with cross-session resume via state tracking |
| ship | Commit, create a PR, monitor CI, and merge a completed phase |
| implement-ship | Implement a phase and ship it (commit, PR, CI, merge) in one session |
| implement-ship-all | Autonomously implement and ship all remaining phases from a plan |
| monitor-pr | Background CI monitoring with auto-merge — checks status every 4 minutes and fixes failures |
Usage (after installation): Start with /plan to design a feature, then /implement to build it phase by phase, and /ship to get it merged. For a fully autonomous flow, use /implement-ship-all (or say "autopilot") to run from plan to merged PRs without stopping.
Domain-agnostic literature management for research projects. Covers OpenAlex search and DOI validation, APA 7 citation generation, LaTeX-safe BibTeX export, full-text acquisition, extraction, citation crawling, and lightweight pipeline orchestration.
| Skill | Description |
|---|---|
| lit-search | Search OpenAlex for candidate papers and inspect normalized metadata |
| lit-validate | Validate a DOI or title against OpenAlex before writing references |
| lit-cite | Add a normalized citation to _index.json and regenerate references.md |
| lit-latex | Export references.bib, lint BibTeX formatting, resolve citekeys, check .tex files, and optionally smoke-test LaTeX compilation |
| lit-acquire | Locate full text via Zotero, OpenAlex OA links, or arXiv |
| lit-extract | Extract structured markdown from PDFs |
| lit-crawl | Expand forward and backward citation chains |
| lit-run | Orchestrate multi-step review pipelines from tracked state |
| lit-status | Inspect pipeline status without side effects |
Usage: Ask Claude to search papers, validate references, generate citations, or maintain a LaTeX bibliography. For BibTeX verification, the plugin now includes offline sample fixtures and scientific-literature/scripts/verify_latex.py.
Install plugins using Claude Code's /install-plugin command. See the Claude Code plugins documentation for the latest syntax.
# Install from GitHub directly
/install-plugin matthiaskloft/claude-skills/bayesflow
/install-plugin matthiaskloft/claude-skills/code-sentinel
/install-plugin matthiaskloft/claude-skills/workflow-automation
/install-plugin matthiaskloft/claude-skills/scientific-literature
Claude Code marketplace entries for the plugin-safe Antigravity Awesome Skills library and its compatible editorial bundles.
Production-ready workflow orchestration with 84 marketplace plugins, 192 local specialized agents, and 156 local skills - optimized for granular installation and minimal token usage
Directory of popular Claude Code extensions including development tools, productivity plugins, and MCP integrations