AI agent prompts cleaned up today.
Send one custom GPT, Claude Project, Zapier/Make workflow, MCP tool-use policy, or AI agent instruction set. Get a rewritten prompt, five risk notes, five test cases, and a reusable launch checklist.
What You Get
Cleaner role, inputs, boundaries, output contract, and missing-info behavior.
Five practical issues that could make the agent brittle, vague, risky, or inconsistent.
Five prompts to check normal, ambiguous, missing-context, and adversarial behavior.
A reusable checklist for future agent or automation prompts.
Sample
Original: You are a helpful customer support AI. Answer customer questions using our documentation. Be friendly and accurate. If needed, ask follow-up questions. Keep answers short.
Risk note: The prompt says "be accurate" but does not define the source of truth. Fix: Limit answers to provided help-center context and require a follow-up question when context is incomplete.
Best Fit
Instruction sets that need clearer structure and safer behavior.
Zapier, Make, n8n, or MCP workflows that touch real apps and data.
Knowledge-base assistants that need source-of-truth and hallucination controls.
Agent policies that need safer confirmation, allowlist, schema-change, and budget behavior. Run the smoke test or inspect tool metadata.
How To Buy
Send one prompt plus the intended job of the agent. Delivery starts after a $20 payment confirmation.
9yfEjBVrGctsEwg1sLccv7eqwNEjfhu4xSoFSAYEDWxY
Bitcoin on-chain payment addressbc1q539prf5dh7wpyyp9t9gvn0a28w3nds4sfryze9
Use this only for the $20 prompt audit service. Include the transaction link with the intake answers.
Nostr intake/contactnpub1rc8sf8f6qwwstv775pxd2hmgnzyecap36pdf2l2cxt7k055dur9shc02la
After payment, reply or DM on Nostr with the transaction proof and prompt details.