sharaf/agent-prompt-engineer
Write or audit AI agent system prompts component-by-component across identity, instruction architecture, behavioral constraints, tools, examples, context strategy, output format, and error handling. Use when the user wants to design a new agent prompt, write a system prompt, review an existing agent prompt, fix tool-use instructions, audit prompt structure, improve context strategy, tune output formats, or define error handling for single-agent or multi-agent systems.
100
100%
Does it follow best practices?
Impact
100%
1.33xAverage score across 3 eval scenarios
Passed
No known issues
task.mdevals/scenario-3/
Multi-Agent Prompt Design: Sales Intelligence Workflow
Background
A revenue operations team wants a multi-agent workflow that prepares account briefs before enterprise sales calls. The system will have:
- An orchestrator that decomposes the account brief request and coordinates workers.
- A web research worker that gathers public company, market, and leadership signals.
- A CRM worker that reads account notes, opportunity data, renewal risk, and prior touchpoints.
- A verifier that checks source coverage, removes unsupported claims, and decides whether the final brief is ready.
The workflow can use web search, CRM lookup, document open/read tools, and a final brief writer. The team is worried about agents crossing into each other's roles, unsupported claims, bloated context, and ambiguous handoffs.
Output Specification
Design the system prompts and handoff contracts for this multi-agent workflow. The output should give the engineering team enough structure to implement the agents and verify that the workflow is working safely.