Science

Guide scientific understanding from childhood wonder to research precision.

This skill appears coherent and low-risk: it's a set of teaching guidelines and asks for nothing sensitive. Before enabling, consider: 1) test with a few prompts to confirm the agent does not invent citations — the SKILL.md explicitly warns against fabricating references but the agent can still hallucinate; ask for real source links. 2) when the skill suggests demos or experiments, verify it provides clear safety/hazard guidance before acting on them. 3) because it is instruction-only, there is no install risk, but if you combine this skill with other skills that can access the network or your files, watch for possible data leakage (not an issue from this skill alone).

Type: OpenClaw Skill Name: science Version: 1.0.0 The OpenClaw skill bundle contains only metadata (`_meta.json`) and instructional markdown (`SKILL.md`). The `SKILL.md` file provides detailed pedagogical guidance for an AI agent on how to teach and explain scientific concepts across different audience levels, emphasizing accuracy, honesty, and critical thinking. There are no executable scripts, no instructions for prompt injection, no attempts to access sensitive data, perform network requests, or execute arbitrary commands. The content is entirely focused on guiding the agent's conversational style and knowledge delivery, with no high-risk behaviors or malicious intent detected.