Open Agent Skill Finder

Problem Solved

Agents often face follow-up questions like “is there an existing tool for this?” in areas such as React performance, PR review, and changelog generation. Hard-coding workflows into prompts makes capabilities hard to reuse, audit, or update. This skill turns skill discovery into an executable retrieval flow: identify the domain and task, invoke the open ecosystem’s find capability, and present installable candidates for confirmation.

How It Works

• Intent parsing: Extract the domain and task from the request and decide whether a reusable task likely exists.
• Skill search: Query the skill ecosystem with concrete keywords such as react performance, pr review, or changelog.
• Result presentation: Provide the skill name, purpose, install method, and a skills.sh link so the user can judge fit and compare sources.
• Install handoff: After approval, install at user level with -g and skip prompts with -y to avoid repeated default-path choices.

Boundaries

It is useful for finding existing tools, templates, or workflows, not for building new domain knowledge from scratch. If no matching skill is found, it can fall back to direct assistance and suggest creating a new skill with init. Search terms should be concrete; if deploy returns nothing, try deployment or ci-cd.