---
name: al-422-when-a-tool-or-platform-trips-up-your-coding-agent-have-t
description: "The agent that just hit a confusing error is the best placed to describe it. On Oct 23, 2026 Vercel added a submit_feedback tool to its MCP server and a vercel feedback command to its CLI, so coding agents can send bug reports, friction reports and feature requests straight to the Vercel team with structured context: the task, the expected behavior, what actually happened and the type of report. It only sends feedback when you ask, and Vercel's docs say to leave out credentials, secrets and full conversations. The prompt below turns that into a habit for any platform you use, with you approving every report. Use when the user wants help with bug reports to vendors (coding) for a small business."
---

# When a tool or platform trips up your coding agent, have the agent send the vendor a clear bug report with the task, what it expected and what actually happened

The agent that just hit a confusing error is the best placed to describe it. On Oct 23, 2026 Vercel added a submit_feedback tool to its MCP server and a vercel feedback command to its CLI, so coding agents can send bug reports, friction reports and feature requests straight to the Vercel team with structured context: the task, the expected behavior, what actually happened and the type of report. It only sends feedback when you ask, and Vercel's docs say to leave out credentials, secrets and full conversations. The prompt below turns that into a habit for any platform you use, with you approving every report.

## Inputs to collect first
Ask the user for anything below that they haven't already given you. Don't guess missing facts.

- Vercel, Supabase, Stripe, other

## Instructions

Do the task below with the user's inputs in place of each [bracket].

While you work on this project, keep a short list of any time a tool, API or platform we use ([Vercel, Supabase, Stripe, other]) blocked or confused you: unclear errors, wrong docs, missing features. At the end of the session, show me each one as a draft report with: the product, what you were trying to do, what you expected, what actually happened (the exact error message), the command or tool you used, and whether it's a bug, friction or a feature request. Remove any secrets, keys, customer data and private code before showing me. Don't send anything yourself. For each report I approve, send it through the vendor's official channel: for Vercel, the submit_feedback tool in the Vercel MCP or `vercel feedback "message" -f task=... -f expected=... -f actual=... -f type=bug`; for others, give me the issue text and the link to their feedback or issue page. Never send the same report twice.

## Safety checks (share these with the user)

- Remove names, contact details, account numbers and other personal data the task doesn't need before you paste, and check your AI tool's data and training settings.
- Read and edit the draft before it reaches a customer or goes public. If you later automate this prompt, start with "draft and notify me", not "send automatically".
- Check names, dates, numbers and any claim against your own source material before you use the answer. AI can misread or invent details.
- Say clearly which parts need a person to review or approve before anything is sent, paid, posted or filed.

## About this skill

- Works with: Vercel MCP, Vercel CLI, Claude Code, Cursor, Codex
- Based on this real-world example: https://vercel.com/changelog/send-feedback-from-vercel-mcp
- Free prompt from AgentPlaybook Real Estate: https://agentic-library.netlify.app/prompt/al-422-when-a-tool-or-platform-trips-up-your-coding-agent-have-the/
- To install in an agent that supports Agent Skills, save this file as SKILL.md in a folder with the same name as `name` above (for example in your skills directory). Install steps for Claude.ai, Claude Code, Cursor and Codex: https://agentic-library.netlify.app/skills/