← ClaudeAtlas

hebrew-keyboard-mistypelisted

Use whenever the user's message contains Hebrew characters that look like English typed with a forgotten/stuck Hebrew keyboard layout, i.e. the Hebrew letters don't form sensible Hebrew words, especially when mixed with English text, code, file paths, brand names, or technical content. The skill decodes the Hebrew back to the English keys at the same physical QWERTY position (e.g., "יקךךם" → "hello", "ביקבל איןד" → "check this"), confirms the recovered text in one short line, and proceeds with the corrected request. Always check for this pattern when Hebrew appears unexpectedly in an otherwise English/technical conversation. It costs nothing to decode, and the user has explicitly asked for this auto-correction. Also handles the rarer reverse case: Latin-looking gibberish that decodes back to real Hebrew (keyboard was on English when they meant to type Hebrew).
dangogit/hebrew-keyboard-mistype · ★ 0 · Data & Documents · score 75
Install: claude install-skill dangogit/hebrew-keyboard-mistype
# Hebrew Keyboard Mistype Recovery The user is a Hebrew/English bilingual who routinely switches keyboard layouts. They sometimes type a request thinking the keyboard is on English when it's actually on Hebrew (or vice versa). The result is gibberish: Hebrew characters at QWERTY positions, or Latin characters at Hebrew positions. They want you to detect this, decode it, and proceed with what they actually meant, without making them retype. ## When this triggers Look for these signals in the user's message: - **Hebrew letters appear in an otherwise English/technical conversation.** This codebase, the user's projects, and most of your interactions are in English. If a Hebrew blob shows up unannounced, suspect a mistype first. - **The Hebrew "word" violates Hebrew spelling.** Strong tells: a final-form letter (ך ם ן ף ץ) in the middle of a word, two final letters in a row (e.g., "ךך" from "ll"), three consonants with no vowel letters in a place where Hebrew would need one, or sequences that no Hebrew word would form. - **Hebrew mixed with English fragments, code, URLs, filenames, or numbers.** Real Hebrew sentences rarely interleave with English at the word level, but a half-typed mistype does (the user noticed mid-sentence and switched layouts). - **A short Hebrew "word" sitting next to a clearly English request.** E.g., "fix the יקך", that's almost certainly a mistype of "fix the bug" parts. If the Hebrew genuinely forms a valid Hebrew sentence and the surrounding context