Oriyn Skill

Install oriyn-ai/skills so agents know when to use Oriyn, how to run readiness checks, and how to summarize research results without overstating confidence.

For direct agent access, connect Oriyn MCP first. The MCP server exposes authenticated tools. The skill gives agents the operating instructions for when and how to use those tools or the CLI.

Install The Skill

npx skills add oriyn-ai/skills

The skill is separate from MCP and the CLI. MCP gives agents direct tools, the CLI performs authenticated terminal work, and the skill tells an agent when to reach for Oriyn and how to present the result.

When To Use It

• A user asks whether to build, price, launch, or redesign a product experience.
• A PR changes user-facing behavior and needs grounded product signal.
• The agent needs actual personas, patterns, or research context for the linked product.
• The agent has Oriyn MCP connected and needs guidance for which tools to call.
• A roadmap decision needs a fast pre-production read before deeper validation.

Skill Workflow

1. Prefer Oriyn MCP when it is connected; use the CLI in terminal or CI workflows.
2. Run oriyn status and confirm auth, API reachability, and product linkage when using the CLI.
3. Inspect personas or patterns if the decision needs context.
4. Run the relevant research workflow for the concrete question.
5. Summarize the structured result, participant summaries, and caveats.
6. Do not claim the result proves real-world impact; frame it as decision support.

CI And Sandboxed Agents

export ORIYN_ACCESS_TOKEN=<token>
oriyn status
oriyn research ab-test --question "<question>" --a "<variant A>" --b "<variant B>" --json

Commit oriyn.json so the agent does not need to infer which product a repository belongs to.