hebrew-nikud

What This Skill Does

The hebrew-nikud skill serves as a highly specialized reference module for OpenClaw AI agents, providing standardized rules for the application of Hebrew diacritics (nikud). In modern Hebrew text-to-speech (TTS) systems, the ambiguity of unvocalized text often leads to mispronunciation. This skill provides the logical framework to determine when and where to place specific vowel points (patach, kamatz, shva, etc.) and dagesh markers. By adhering to the golden rule of 'only adding nikud when 100% certain,' this skill prevents common TTS errors that arise from guessing word structures based on context alone.

Installation

To integrate this skill into your local OpenClaw environment, execute the following command in your terminal: clawhub install openclaw/skills/skills/shaharsha/hebrew-nikud

Use Cases

This skill is essential for developers and content creators working with Hebrew language AI applications. Primary use cases include:

1. Preparing text strings for high-fidelity TTS engines where pronunciation of ambiguous consonants (begedkefet) is critical.
2. Distinguishing between homographs (words spelled identically but pronounced differently) that confuse machine learning models.
3. Managing gender-specific suffixes that alter vowel structure, ensuring the AI output correctly reflects the speaker's intended gender context.
4. Standardizing the usage of the Dagesh Forte (doubling) in verb patterns like Pi'el or Hitpa'el.
5. Assisting in the correct pronunciation of loanwords and foreign names that do not follow standard Hebrew root patterns.

Example Prompts

1. "Apply selective nikud to this sentence to ensure the TTS reads the homograph correctly: 'הַסּוּס רָץ בַּשָּׂדֶה' - ensure the dagesh in the Bet is correct."
2. "Review this verb conjugation for the root P-A-L and add nikud to the Pi'el present tense form for correct articulation."
3. "Which Shva rule applies to the word 'Bereshit'? Explain and provide the properly nikud-marked version."

Tips & Limitations

Always remember that in Hebrew, less is often more. The primary limitation of this skill is that it is not intended for full-text vocalization, which can be computationally heavy and prone to error. Instead, use it surgically to clarify points of confusion. Note that dagesh rules regarding the letters Gimmel, Dalet, and Tav are considered vestigial in Modern Hebrew and generally do not impact TTS performance; focus your efforts on the Bet, Kaf, and Pe (the audible begedkefet letters) to yield the highest quality audio output.