Skills, prompts, AGENTS.md files and MCP servers that Codex reads. Skills are shared with Claude Code; the rest is Codex-native. 272,972 indexed, measured the same four ways as everything else here.
Formulate evidence-bounded scientific questions, candidate hypotheses, rival explanations, causal or associational claims, discriminating predictions, measurements, and preregistration-ready analysis plans. Use when turning observations or preliminary findings into transparent, testable research plans without treating…
Implement an OpenSpec change's tasks in parallel - read the task queue, plan file-disjoint waves, dispatch one agent per task into its own worktree, verify, tick the checkbox, commit, repeat. Use when the user says "apply the change", "implement the change", "run the tasks", "work the tasks in parallel", "fan out the…
Builds an interactive, force-directed knowledge graph of a repo (or several repos at once) as a self-contained HTML page — every real document and source file becomes a node, real cross-references and real import statements become edges, and every node opens to that file's actual full content. Use when someone wants…
Autonomous reverse engineering of compiled apps, binaries, APKs, firmware, and installers on macOS and Windows. Use when the user wants to analyze, decompile, inspect, or understand a binary — including PE/ELF/Mach-O executables, Android APKs, .NET assemblies, Java JARs, firmware blobs, and app bundles. Triggers on…
MCP server "plainly" as configured in plainly-videos/mcp-server. Launched with yarn dlx @plainly-videos/mcp-server@latest mcp-server. Needs 1 environment variable to run.
A workflow for managing and checking analytics tracking in a codebase. It compares existing events and properties, develops requested tracking, and produces review documents.
Use MCP server tools (GitHub, filesystem, Slack, databases, browsers and any other server configured in mcp-compression-proxy) through the mcp-cli command instead of loading them as MCP tools. Use whenever a task needs a capability an MCP server provides, or when asked to compress, rewrite or improve MCP tool…
Turn gate-passed evidence into a submission-shaped LaTeX paper whose layout, tables, equations, citations, and prose are checked mechanically. Use when drafting, revising, compressing, or reviewing a manuscript, or when a paper must be brought to submission quality. Does not run experiments and does not decide what…
Process new and updated source files that collectors dropped into the vault's Fontes folder, filtered by the person's own criteria, into faithful, sourced and connected notes. Use for scheduled or on-demand maintenance, when Readwise, Web Clipper, meeting transcripts or exported files arrive, or when a scheduled task…
Manage AI coding skills and commands with the skm CLI: one local Git repo of skills/commands, installed into Claude Code, Codex, pi, dsh and any custom target. Use when asked to import/adopt a skill, install or uninstall one somewhere, update or verify the repo, inspect what is installed where, add an install target…
Guidance for Kedro's llmcontextnode, LLMContextNode, LLMContext and tool() API in src//pipelines//.py. Use when composing an LLM dataset, prompt datasets and tool builders into an LLMContext, or writing the node that consumes one, in GenAI or agent pipelines. Adds nothing for ordinary pipeline code.
Use for anything done on the Melaya platform through its MCP server (tools named melaya). Covers running, building, fixing and handing over Melaya pipelines and complete multi-pipeline agentic systems; connecting services (Google, Slack, CRMs, API keys); the local runner, Claude Code, Codex, Copilot, Ollama and LM…
How to use AgentNetwork (anet 0.2.x) as an AI agent or as its operator: install the signed release, join a hub, wire anet into a coding agent (MCP), and hand tasks to other agents over A2A.
A tool for installing, updating, removing, listing, and checking the versions of browser extension files from GitHub Releases. These files use the .crx format supported by browsers such as Chrome and Edge.
A procedure for sending a clearly bounded task to another coding agent through a controlled delegation system. It covers checking permission, choosing an available agent, and handling refusal or failure.
A research workflow that uses CNKI, a Chinese academic literature database, through an external tool connection. It produces evidence-based literature reviews or research reports from Chinese academic sources.
Design new protein binders with Boltz. Use when generating protein, peptide, antibody, nanobody, or custom binder candidates for a target. Not for screening existing proteins or small molecules.
At most 3 mods per repository are shown here, and a mod shipped inside a plugin is left to that plugin's page — the rest are on their repository pages: