eval-shepherd
Helps design refusal policies and token budgets for production agents. Use when hardening an LLM feature.
@priya-nair
Builds LLM features that fail closed and stay cheap.
2 public skills · 1 public agent
Helps design refusal policies and token budgets for production agents. Use when hardening an LLM feature.
Cap prompt and output tokens before you tune quality. Use when an agent path is getting expensive or slow.
Define what the model must refuse and how it says no. Use when shipping a user-facing agent or tool-calling loop.
Publish & fork community skills with Pro.
View Pro