extractlisted

# Extract (Codex variant) Pull content out of every URL in a notes file (YouTube subtitles, Telegram post text, HTML articles) into a shared sibling `extracted/<note-basename>/` directory, replacing each URL with a local pointer. Processing multiple notes in the same directory consolidates under one `extracted/` parent. This is the **Codex CLI variant**. Behaviourally identical to the Claude Code variant — only the user-interaction idiom differs (numbered list TUI prompt instead of `AskUserQuestion` radio UI). All shared resources (`roles/`, `scripts/`) come from the Claude variant tree via install-time symlinks. > **Letter = spirit.** If a rule blocks you from reaching the goal it was > written for, the rule is wrong, not the goal. Don't look for a wording > loophole — ask what the rule is protecting, and protect that. ## Usage ``` /extract <note.md> [--force] [--light | --full] ``` `--force` re-processes URLs even if they're already annotated (default: skip already-annotated). `--light` writes a one-line summary of each link's gist **inline** next to the URL (no `extracted/` tree); `--full` (default) extracts full content into `extracted/`. If neither flag is given, the skill asks which mode at the start (step 0, numbered TUI prompt). ## Weaknesses and when NOT to use - **Does not work with private/auth resources.** Private Telegram channels, paywalled articles, logged-in-only pages — out of scope. Those URLs return an error in the final report. - **Depends on exte