Herramientas

Open-source LLM evaluation platform: one opencompass command picks the model and datasets and runs the evaluation across hundreds of benchmarks. Works with local Hugging Face models and API models, with vLLM and LMDeploy for faster inference.

Converts and two-way syncs Jupyter notebooks with plain Python scripts or Markdown, so agents can edit research and teaching notebooks like ordinary text and diffs stay readable. Needs a local Python environment; install with pip.

Cleans a LaTeX project before arXiv submission: removes comments, auxiliary files and unused .tex files and images, and can compress images, so internal notes are not published with the source. From Google Research; installable with Homebrew on macOS.

Python library and CLI for extracting main web content: fetches pages, crawls via sitemaps and RSS feeds, strips headers, footers and other noise, and outputs TXT, Markdown, JSON, XML and more with metadata. Also processes downloaded HTML offline.

Anthropic's official OS-level sandbox CLI: restricts file and network access for agents, MCP servers and any command without containers, using sandbox-exec on macOS and bubblewrap on Linux. Currently a research preview.

Static security analysis for CI/CD config: finds template injection, credential leaks, excessive permissions and more in GitHub Actions workflows, also covers Dependabot and pre-commit configs, and auto-fixes some findings.

Lets agents launch the Godot editor, run projects in debug mode and read console output and errors, plus create scenes, add nodes and manage UIDs (Godot 4.4+). Requires Godot and Node.js installed locally.

Veteran suite of CSV command-line tools: in2csv converts Excel and JSON to CSV, csvcut, csvgrep and csvsort filter and sort, csvstat computes statistics, and csvsql runs SQL directly against CSV files.

Google's official skills for building ADK agents: dev workflow, scaffolding, the ADK Python API, evaluation with LLM judges and adaptive rubrics, deployment to Agent Runtime, Cloud Run or GKE, publishing and observability. Works with agents-cli; local dev needs only an AI Studio key.

250+ MCP tools for Ghidra: decompilation, cross-references, data-flow analysis, bulk renaming and commenting, P-code emulation and live debugging, with headless mode support. Requires Java 21 and Ghidra 12.1, and the plugin must be built and deployed first.

Semi-automated research workflow for computer science and AI researchers: ideation, Zotero literature management, experiment analysis and reports, paper writing and polishing, and rebuttals, with optional Obsidian knowledge base. 40+ skills; also installable as a Claude Code plugin.

Semi-automated research assistant for CS and AI researchers that links ideation, Zotero literature management, experiment analysis and reporting, paper writing and reviewer responses, and keeps project knowledge in Obsidian. Branches for Codex and other agents exist.

LM Studio's official CLI: list, load and unload local models, start and stop the local API server, and check status and logs; ls, ps and other commands support --json output for scripts and agents. Requires the LM Studio desktop app.

Detects shot changes in video and splits long footage into scenes, exporting cut lists and per-scene thumbnails, useful for shot analysis, organizing footage and finding edit points. Runs locally; needs Python 3.10+, plus FFmpeg or mkvmerge for splitting.

Classic CLI for reading and writing metadata in hundreds of file types including images, video, audio and PDF, covering EXIF, GPS, IPTC, XMP and camera makers' proprietary tags, with batch editing, stripping and conversion. Requires Perl.

CLI and TypeScript runtime for discovering and calling MCP servers from the terminal and scripts: list tool signatures and call tools directly, or generate a standalone CLI or typed client from a server. Auto-imports existing MCP configs from Claude Code, Codex and others.

Draws bar charts, histograms, line charts, scatter plots and box plots in the terminal from CSV/TSV on stdin, so you can glance at data distributions at the end of a shell pipeline. The command is uplot.

Teaches agents to code against the latest Gemini API and Python/TypeScript SDKs: multi-turn chat, function calling, structured output, and image and speech generation, while avoiding deprecated older models. Maintained by Google's Gemini team.

Gemini API development skills: gemini-api-dev (core API and SDKs), gemini-live-api-dev (real-time bidirectional audio/video streaming) and gemini-omni-flash-api (Omni video generation and editing), filling in APIs released after model training cutoffs. Maintained by Google's Gemini team.

Three.js texture reference with code: loading and configuring texture maps, UV mapping, cubemaps and HDR environment maps, texture settings and optimization.

Three.js geometry reference with code: parameters for built-in shapes like boxes, spheres, cylinders and tori, building custom meshes from vertices with BufferGeometry, and instanced rendering for large numbers of objects.

Three.js lighting reference with code: ambient, hemisphere, directional, point, spot and area lights with their cost and shadow support, shadow setup, image-based environment lighting and lighting performance tips.

Three.js materials reference with code: when to use Basic, Lambert, Phong, Standard, Physical, Toon and other materials, PBR properties such as roughness and metalness, custom ShaderMaterial and material performance.

Three.js post-processing reference with code: chaining passes with EffectComposer, adding bloom, depth of field, color grading, blur and glow, and writing custom screen-space shader passes.