Prompts
Showing 145 to 156 of 578 results
Work Out Why an Agent Run Went Wrong
Reads an agent trace to find the first step where the run actually diverged — usually well before the visible failure — classifies the failure type...
Design an Agent's Memory Across Sessions
Decides what an agent should remember between sessions, what it must forget, how memories are retrieved without flooding the context, and how stale...
Split Work Between an Orchestrator and Subagents
Decides whether a task genuinely needs multiple agents, then draws the boundaries so each subagent has an isolated context and a verifiable contrac...
Handle Errors and Retries Inside an Agent Loop
Designs the failure behaviour of an agent loop: which errors are worth retrying, how to keep failed attempts from poisoning the context, and when t...
Design the Tool Surface for an Agent or MCP Server
Decides which tools should exist and at what granularity, names them so the model picks correctly, and prunes the surface — because tool-selection ...
Harden a Prompt or Agent Against Prompt Injection
Maps where untrusted text reaches your model, what an attacker could reach through it, and which mitigations are architectural versus merely hopefu...
Scope What an Agent Is Allowed to Do
Defines an agent's permission surface by blast radius rather than convenience: what it can do freely, what needs approval, what it must never do, a...
Choose Few-Shot Examples That Actually Help
Selects, orders, and stress-tests few-shot examples so they teach the boundaries of a task rather than one happy path — including when few-shot is ...
Design Retrieval That Finds the Right Thing
Designs a retrieval layer around the questions it actually has to answer — chunking, what gets indexed, filtering, reranking — and insists on evalu...
Budget an Agent's Context Window
Treats context as a budget to spend rather than a container to fill, allocating tokens across system prompt, tools, history, and retrieved content,...
Choose the Right Model for a Task
Works out which model tier a task actually needs by defining the hardest case it must handle, then designs the head-to-head test that settles it on...
Decide Whether Your Task Actually Needs an Agent
Tests a task against the compounding-reliability math before you build an agent for it, and lays out the cheaper architectures — single call, fixed...