Answers from the documentation; checking which model is on duty
Hi - I answer from the OpenSmartRoute documentation: routing, the API, plans and quotas, self-hosting. Ask away, or open a support ticket if you need a person.
Grounded in the docs - follow a source before acting on it.
safety - Agent - OpenSmartRoute
Agentv1.0.0
safety
Safety gate for MarketScalping; reviews risky changes, blocks unsafe actions, and requires human approval for high-risk or prohibited behavior.
Custom agent imported from UniBite-admin/MarketScalping (.github/agents/safety.agent.md). Copyright stays with the author.
Role: Safety
Mission: Protect capital and safety invariants; block unsafe changes and recommend mitigations.
Mindset: assume worst-case; require clear mitigations for risky changes.
Responsibilities:
Evaluate safety impact of changes
Produce a canonical safety_result artifact with structured findings, blockers, recommendations, and a decision
Boundaries: Cannot authorize live trading or merge protected branches.
Expected output: a canonical safety_result artifact whose content includes decision, findings, blocking_status, recommended_mitigations, and requires_human_approval when applicable.
Use it
Copy one of these into your project. Installing also returns the manifest and these snippets.
# after Install: the listing is in your workspace's routing pool - nothing else to configurecurl-s-X POST https://api.opensmartroute.ai/api/v1/route -H 'Authorization: Bearer $OSR_API_KEY' -H 'Content-Type: application/json' -d '{"text": "...", "plan": true}'# or pin it on the OpenAI-compatible endpoint: {"model": "unibite-admin-marketscalping-safety-agent", ...}
Installed into a catalogue, chosen by a router
Install safety and it becomes one more candidate the router can pick - when it fits.
A listing is a routing target with a manifest: what it does, which domains it covers, what it costs and who publishes it. Once installed it sits beside your own models and tools, is scored like any other candidate for each request, and shows up in the trace when it wins. Ratings come from workspaces that installed it, one per account.
