# Your Skill Name

> What this skill does in one sentence. Trigger when the user asks the agent to do X, review Y, or check Z.

- Skill: `flowleap-ai/your-skill-name` (Agent Skill)
- Install (CLI): `npx skillmds@latest add flowleap-ai/your-skill-name`
- Raw SKILL.md: https://api.skillmd.com/api/skills/flowleap-ai/your-skill-name/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: AI & ML
- Author: flowleap-ai (https://skillmd.com/u/flowleap-ai)
- Updated: 2026-09-17
- Page: https://skillmd.com/skills/flowleap-ai/your-skill-name

---


# Your Skill Name

One or two sentences of context: who is using this and what they are trying to
achieve. Write the body as instructions addressed to the agent.

## How to do it

Break the task into clear steps. Every command must call the **FlowLeap CLI**
(`flowleap …`) — never the FlowLeap app's internal typed tool names like
`search_patents(...)`. Run `flowleap --help` to see the real commands, and only
put in commands you have actually run.

```bash
# Example — replace with the real commands your workflow needs:
flowleap patent build-query "wireless charging for electric vehicles"
flowleap patent search --query "ti=wireless AND ti=charging" --limit 20
flowleap ops claims EP3456789
```

Done when <a clear, checkable finish line — e.g. "every live hit has its claims
pulled and a legal-status verdict">. Write a finish line the agent can verify,
not a fuzzy one like "when you have enough results."

## Optional: pointing at other skills

If a deeper workflow lives in another skill, mention it as a suggestion that still
makes sense when that skill is NOT installed — for example: "If the full recipe
pack is installed, `recipe-prior-art-search` runs this end to end." Never write a
step that dead-ends when the other skill is missing.