Policy Explanation
Generating plain-English description…
What this policy does
Asking the AI to explain this policy…
Generating plain-English description…
Plain-English summary of the active rules — what agents can do, what gets blocked, and what needs a human.
In the dev environment, the **research_crew** agent may use only the **arxiv_search**, **web_search**, and **delete_research_data** tools, and it may run only on **gpt-4o-mini**. Any use of **delete_research_data** requires approval first. Requests are limited to **4,000 tokens per request**; usage is monitored with an **80% warning threshold**, and excess use is tightly controlled with a **24-hour UTC window**, a **hard deny at 2× the overage limit**, and a **repeat limit of 3**. Safety guardrails require **PII to be redacted**, and any **prompt injection** or **data exfiltration** attempts must be **denied**.