codex-skills

Use when a task explicitly involves Microsoft MarkItDown/markitdown-mcp, or deciding whether to use MarkItDown to convert documents or URIs into Markdown for ingestion. Covers MarkItDown-specific extras, MCP behavior, plugins/OCR/Azure options, security boundaries, and format limitations; avoid for high-fidelity Office/PDF editing.

Prepare, verify, start, stop, and troubleshoot a safe reverse SSH workflow where a macOS machine dials out to a VPS and exposes only the Mac's localhost SSH service to the VPS loopback address. Use when Codex needs to help Tuzi/Hermes or another remote agent access a user's Mac without opening public router ports, including Remote Login checks, authorized_keys setup, reverse tunnel commands, autossh/launchd drafts, network/VPN diagnosis, and maintenance after reboot or network changes.

Review current code changes for reuse, quality, and efficiency, then write a simplify implementation plan instead of modifying code. Use when the user wants simplify-style investigation first but explicitly wants a plan, a plan-only pass, a planning mode run, or a markdown implementation plan in the current workspace rather than immediate code changes. In Codex, always launch three parallel explorer subagents with the full diff, using model gpt-5.4 and reasoning_effort xhigh, then synthesize their findings into a single plan document.

Retrieve web page content in clean markdown/text for research and summarization. Use when Codex needs to read a URL, bypass noisy HTML, or handle Cloudflare-prone pages. Always use this fallback chain in order: markdown.new, then defuddle.md, then r.jina.ai, then Scrapling. Persist markdown/image assets under the current Codex session directory, and when image understanding is needed, read downloaded local images first.

Review current code changes for reuse, quality, and efficiency, then fix the worthwhile issues without changing behavior. Use when the user asks to simplify, clean up, de-duplicate, polish, or make recent edits more maintainable, especially after a coding pass or before commit or review. In Codex, follow the same three-agent simplify workflow as the source simplify prompt: always launch three parallel explorer subagents with the full diff, using model gpt-5.4 and reasoning_effort xhigh, then aggregate findings and fix the worthwhile issues in the main thread.

create or refactor repository documentation and agents files using progressive disclosure. use when chatgpt needs to add new docs, split oversized docs, build indexes, or maintain readme and agents structures so coding agents can navigate a codebase without loading everything at once. optimized for repositories that want a hard maximum of three layers, explicit read triggers, compact routing docs, and leaf docs that contain the detailed implementation facts.

Install, configure, repair, update, or rebind the Telegram Codex Bridge after the user installs this skill. Use when the user wants Codex to take over bridge setup with minimal user action, only interrupting for unavoidable external steps like providing a Telegram bot token or messaging the bot once.