Skills, prompts, AGENTS.md files and MCP servers that Codex reads. Skills are shared with Claude Code; the rest is Codex-native. 275,525 indexed, measured the same four ways as everything else here.
Karthik's single-equity investment framework for 5–10-year holders. Turns a ticker into a decision (Initiate / Add / Hold / Trim / Exit / Watch / Avoid), tested against a fixed 13% annual hurdle on a probability-weighted five-year value built from unit economics, ROIC, owner earnings and a moat read. Use whenever the…
Taiwan Traditional Chinese (zh-Hant-TW) house style: terminology, punctuation and de-AI tone. Use it when your output contains Chinese. Decide by the output, not by keywords in the task. Read it before you write the first Chinese character. Git, React and JSON tasks are included when the output is Chinese. Examples…
Convert PowerPoint (.pptx) decks into Slidev Markdown that looks like the original, checked against a reference PDF (click steps, crops, lost words, compare images, fresh reviewer) with image provenance; also re-checks earlier conversions. Use when the user wants a .pptx (or "this week's lecture") as Slidev or…
Autonomous drupal.org issue fixer called LoachBot. Picks up one open drupal.org GitLab issue, claims it by commenting /do:assign me, implements the fix on the issue's shared fork, opens a merge request, then unassigns itself and sets state::needsReview with a short summary. Use when the user wants to pick up or fix a…
Use when the user asks about cloud, Kubernetes, SaaS or AI spend through Costory: why a bill moved, cost by team/service/namespace, unit economics, budgets, allocating shared or untagged cost, or building Costory dashboards, scheduled reports and cost alerts. Covers org selection, the tool order, preview-then-confirm…
★not rated 2
changed 3d agoA91 tokens
originalApache-2.0
Restart this Claude session inside the git repository it is working on, keeping the conversation. Use when a remote (crctl / screen) session was started outside its repo - e.g. in the home directory - so the Claude apps show no branch, diff or pull request for it; or when the user asks to restart this session, move…
Create a bounded handoff for another agent or person on the board with handoffsend, after confirming the recipient, the request and the sensitivity with the user.
Export GeoGebra diagrams to editable TikZ, Asymptote, or TSQX, preserving construction relations and reconstructing unsupported definitions with AI assistance.
A tool for creating, reusing, and editing 1C configuration objects, managed forms, and registered table-template content in XML source files. It works through a metadata command-line tool, which is a text-based program for manipulating this configuration data.
Use this skill for market research or trading questions when the Slatemark tools are connected. Triggers include: market-data analysis, position review, trade-idea evaluation, portfolio questions, options analysis, macro setup checks, trade journaling (logging position intent, user-reported activity, or an outcome)…
Build slide decks (PDF, PPTX), A4 reports (PDF, DOCX), Excel workbooks and draw.io figures, search papers and the web, and read existing PDF/DOCX/Markdown files, with the CREPE MCP tools. Use for any deck, report, workbook, diagram, literature search or file to read.
Find current Tokyo nightlife events, venues, performers, guest lists, and VIP table options using the Nightlife MCP. Use when someone asks what is on tonight or this weekend, where to go out in Tokyo, who is playing, or wants help joining a guest list or requesting a VIP table.
Use when the user asks to find, create or update test cases in Allure TestOps, add steps and expected results, tag or link test cases, run a manual launch, or triage failed test results and defects. Triggers include "create a test case in Allure", "find test cases for this feature", "update the expected result", "what…
Interface with the Veris AI agent simulation and evaluation platform. Use this skill when you need to create AI agents, run simulations, generate scenarios, or evaluate agent performance. The Veris platform provides comprehensive tools for testing and validating AI agents through controlled simulation environments.
MCP server that automates image generation on ChatGPT web, edits images by instruction, and saves them locally. Runs locally from the image-generation-mcp-server npm package.
★not rated 2 6d agoASocket: warn
tokens not measured
originalMIT
Cuts a llama-launcher release: checks the release gate with go run ./cmd/release status, drafts the notes with release-notes-llama-launcher, proposes the version and has the owner confirm it, makes the two local release commits with prepare, then — after a dry run and one explicit confirmation — runs publish, which…
Read-only review of architecture and engineering quality against the repo's canonical shape (owners, boundaries, names, amount of machinery). Run on request, and exactly once on the first non-draft PR head.
Create, audit and clean up a repo's .scratch/ notes bundle (an Open Knowledge Format folder with its own index.md). Use whenever writing a note, plan, ticket draft, PR description, research or reference doc into .scratch/, and when asked to clean up, audit, sort, promote or reorganize .scratch.
MCP server "hermes" as configured in jozrftamson/hermes-figma-prompt-hub. Runs /path/to/hermes-figma-prompt-hub/mcp/server.py with python3. Needs 1 environment variable to run.
Use only when the user explicitly asks for a "Baseline Security" scan or audit of the codebase, or to create Jira stories from a Baseline Security audit. Do not use for a general security scan, vulnerability assessment, penetration test or code review that does not name Baseline Security.
Run a GEO audit with the SV MCP geogptaudit tool (async): how a domain or page is visible in AI-generated answers for specific entities, with citation targets, schema fixes, quick wins and a roadmap. Use when the user asks for a GEO audit or AI-visibility score for named entities.
Record what a just-finished debugging session taught as a numbered card in PATTERNS.md. Use right after a bug is fixed and its root cause is known, or when the user asks to record, document, or register a pattern or incident.
★not rated 2 3d agoA53 tokens
originalMIT
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: