Item detail
github.com

YutoTerashima/agent-safety-eval-lab

YutoTerashima/agent-safety-eval-lab is a evaluation lab in RepoRadar's Developer Tools section, holding Gold tier and a 'try now' verdict. Its strongest signal is workflow potential, scored 9.8 out of 10.

Score8.3
Popularity1.0
Riskconditional
TierGold
Score breakdown
Usefulness8.0
Novelty8.0
Momentum6.0
Maturity6.5
Open-source/build8.4
Evidence8.0
Workflow potential9.8
Setup ease6.4

Popularity is tracked separately. Support, ads, sponsorships, and tips never affect these signals.

Why it matters

Useful for teams that need to test whole agent trajectories instead of only grading final answers, especially when tool calls and policy boundaries are the real failure surface.

Who should use it

Evaluation teams building repeatable tool-use and policy tests for agentsDevelopers who need a trace-first view of agent failures before shipping broader automationResearchers comparing guardrails, adapters, and tool policies across the same casesPlatform engineers who want mock-mode safety tests before connecting live providers

Who should skip it

Skip YutoTerashima/agent-safety-eval-lab unless the captured evidence suggests it solves a problem you are actively working on.

About this signal

YutoTerashima/agent-safety-eval-lab is tracked by RepoRadar as a evaluation lab in the Developer Tools section. It was first seen on 2026-06-28 and last updated on 2026-06-28. The current verdict is 'try now' with a Gold tier and moderate setup difficulty. YutoTerashima/agent-safety-eval-lab leads on workflow potential (9.8) and open-source/build quality (8.4); its lowest signal is momentum (6.0), so factor that in before investing setup time. This page summarizes the evidence RepoRadar has captured from captured source metadata. The score, tier, risk label, and verdict on this page are never influenced by sponsorship, ads, or tips — they reflect only the usefulness, popularity, novelty, momentum, maturity, and evidence signals described in the RepoRadar methodology.

How this item is evaluated

RepoRadar assigned YutoTerashima/agent-safety-eval-lab a composite score of 8.3 out of 10, placing it in the Gold tier. This score combines weighted sub-signals: usefulness (35%), novelty (18%), momentum (14%), maturity (10%), open-source/build quality (7%), evidence quality (6%), workflow potential (6%), and setup ease (4%). Popularity is tracked separately at 1.0 and never affects the composite score or tier. The risk label of 'conditional' reflects inherent user-impacting hazards, not generic novelty. Items with no risk flag may still require normal code review before production use.

Putting this into practice? Read How to vet an AI agent or MCP server before you wire it in for the checklist behind this score.

Risk explanation

The default mock mode is safe for local evaluation, but real adapters and trace imports can pull sensitive prompts or tool logs into the lab, so scrub production data before replaying it; It is an evaluation harness rather than a plug-and-play guardrail, so you still need your own policy thresholds before wiring results into shipping decisions.

Evidence links
Closest alternatives / related signals
agent-evalssafetytool-usepythontracesmit