Feature Request: Clarvia AEO Scoring Integration
Hi AgentOps team,
I'm working on Clarvia (https://clarvia.art), an open-source tool quality scoring platform that evaluates APIs and MCP servers for AEO (Agent Experience Optimization).
The idea
AgentOps already tracks which tools agents call. Clarvia can enrich those sessions with quality metadata — showing not just that an agent called tool X, but how well-designed that tool is for agent use.
Example enrichment:
- Tool called:
brave_search (MCP server)
- AEO Score: 82/100 (AGENT_READY)
- Signal: well-documented, stable, proper error handling ✓
How it would work
import agentops
from clarvia_langchain import CriteriaGate
# Before each tool call, gate on AEO quality
gate = CriteriaGate(min_rating="AGENT_FRIENDLY")
# AgentOps records: tool name + AEO score + outcome
Clarvia capabilities
- REST API:
GET /v1/score?url=<api-url> → instant AEO score
- MCP server:
npx -y clarvia-mcp-server → 15K+ pre-scored tools
- 15,400+ indexed MCP servers with quality scores
- Free tier available
Why this matters for AgentOps users
Agents that use high-AEO tools have fewer errors, better task completion. Adding AEO scores to AgentOps session data would help users correlate tool quality with agent performance.
Interested in exploring an integration? I can open a PR or collaborate on a plugin.
Links:
Feature Request: Clarvia AEO Scoring Integration
Hi AgentOps team,
I'm working on Clarvia (https://clarvia.art), an open-source tool quality scoring platform that evaluates APIs and MCP servers for AEO (Agent Experience Optimization).
The idea
AgentOps already tracks which tools agents call. Clarvia can enrich those sessions with quality metadata — showing not just that an agent called
tool X, but how well-designed that tool is for agent use.Example enrichment:
brave_search(MCP server)How it would work
Clarvia capabilities
GET /v1/score?url=<api-url>→ instant AEO scorenpx -y clarvia-mcp-server→ 15K+ pre-scored toolsWhy this matters for AgentOps users
Agents that use high-AEO tools have fewer errors, better task completion. Adding AEO scores to AgentOps session data would help users correlate tool quality with agent performance.
Interested in exploring an integration? I can open a PR or collaborate on a plugin.
Links: