2 papers
cs.AI2026
The Regression Tax: Decomposing Why Skills Help and Hurt LLM Agents
Darshan Tank, Baran Nama
Adding procedural skills to an LLM agent is typically evaluated by average improvement in task success. However, this metric hides an important cost: skills can also make agents wo…
cs.IR2026
CryptoAnalystBench: Failures in Multi-Tool Long-Form LLM Analysis
Anushri Eswaran, Oleg Golev, Darshan Tank +2
Modern analyst agents must reason over complex, high token inputs, including dozens of retrieved documents, tool outputs, and time sensitive data. While prior work has produced too…