ykdojo/claude-code-tips

Overview

This skill provides a reliable method to fetch Reddit content using the Gemini CLI inside a tmux session when direct WebFetch access is blocked or returns 403 errors. It standardizes a lightweight workflow to send queries, capture output, and verify submission state so you can extract Reddit threads and comments despite network restrictions. The approach is scriptable and intended for repeatable, automated research tasks.

How this skill works

You create a detached tmux session and run the Gemini CLI inside it. Queries are sent to the Gemini process via tmux send-keys, the session output is allowed time to generate, and tmux capture-pane extracts the response text. The method includes a simple visual check to determine whether Enter was actually submitted and a cleanup step to kill the tmux session when finished.

When to use it

• WebFetch or browser-based access to Reddit is blocked, returns 403, or is unreliable.
• You need to programmatically research Reddit threads or comments from a restricted environment.
• Automating repeated Reddit queries in a headless server or CI pipeline.
• When you want a reproducible, scriptable CLI-based fetch flow using Gemini as a proxy.
• Short ad hoc lookups during development when a GUI client is unavailable.

Best practices

• Pick a unique tmux session name and reuse it consistently for the job to avoid collisions.
• Allow sufficient wait time after sending a query (start with 30s, increase up to 90s for complex searches).
• Always capture pane output with a reasonable scrollback range (e.g., -S -500) to include full responses.
• Detect whether Enter was actually sent by checking whether your query appears inside the tmux input box.
• Kill the tmux session after the job to free resources and avoid stray background processes.

Example use cases

• Researching subreddit threads for sentiment or topic summaries when direct scraping is blocked.
• Collecting comment excerpts for a report when API or web access is restricted.
• Embedding Reddit lookups into a larger agentic workflow that uses Gemini as a helper.
• Debugging why a Reddit URL returns 403 by reproducing the fetch from a different environment.
• Running scheduled Reddit queries on a remote server without a browser session.