Tool use / function calling
ShieldAgent
ShieldAgent: Shielding Agents via Verifiable Safety Policy Reasoning
Superseded baseline#48 of 55 most-superseded · first seen Mar 26, 2025
Cited as a baseline — critiqued by newer work, not yet beaten on a benchmark here
1 papers critique it · 0 beat it on benchmarks
What papers say
Verbatim critique sentences, each from a paper that cites ShieldAgent as a baseline.
ShieldAgent and AGrail produce safety judgments via complex reasoning and verification pipelines, incurring high latency that makes them impractical for step-level monitoring of tool invocation in LLM-based agents.
What to use instead
Recent methods in the same sub-problem, not yet superseded in the knowledge base — arXiv benchmark leaders, not vetted production recommendations.