EmbeddingGemma 2: an open, lightweight multimodal embedding model
deepmind.google·6d ago
AI Engineering
Hacker News·Jul 30·5 min read
Agent Skill to Force Docs in ASD-STE100 Simplified Technical English
TL;DR
An agent skill has been developed to ensure that large language models (LLMs) generate documentation in ASD-STE100 Simplified Technical English, a controlled language used in aerospace since 1983. This innovation aims to prevent misinterpretation of instructions by mechanics.
✦ Why It Matters
Implement this agent skill in your LLMs today to enhance documentation clarity and reduce errors.
Key Takeaways
- Install the agent skill for LLMs to enforce Simplified Technical English.
- Expect clearer documentation with fewer ambiguities and shorter sentences.
- Use the skill for various technical writing tasks like error messages.
- Integrate easily into tools like Claude Code and OpenAI Codex.
How It Works
The agent skill applies 53 specific rules from the ASD-STE100 standard to LLM outputs, ensuring compliance with clarity and simplicity requirements. For example, it limits sentence length to a maximum of 20 words and mandates the use of active voice, which helps prevent misinterpretation.