mcp-feedback

mcp-feedback is a cursor rule for Cursor from dragonstylecc/Interactive-Feedback-With-Capture-MCP. It costs 337 tokens per session, scanned A, original, MIT.

Rules for using an interactive feedback tool that pauses to ask the user questions or request feedback during a task.

In plain words
What is it for?
They are for deciding when to ask for clarification, checking feedback before finalizing work, and falling back to another question method when needed.
Why use it?
They define when clarification is required, how to offer choices, and how to handle missing tools, timeouts, and retries.

Cursor rule for Cursor

Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

agentmods
npx agentmods add rules/dragonstylecc/interactive-feedback-with-capture-mcp/mcp-feedback
Clone the repo
git clone --depth 1 https://github.com/dragonstylecc/Interactive-Feedback-With-Capture-MCP

Made for: Cursor.

Wrote this? Show the measurements

A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.

agentmods badge for mcp-feedback

README.md
[![agentmods](https://agentmods.dev/badge/rules/dragonstylecc/interactive-feedback-with-capture-mcp/mcp-feedback.svg)](https://agentmods.dev/rules/dragonstylecc/interactive-feedback-with-capture-mcp/mcp-feedback)
Your own site
<a href="https://agentmods.dev/rules/dragonstylecc/interactive-feedback-with-capture-mcp/mcp-feedback"><img src="https://agentmods.dev/badge/rules/dragonstylecc/interactive-feedback-with-capture-mcp/mcp-feedback.svg" alt="Measured on agentmods" height="20"></a>
Per session 337 This file is loaded in full into every session.
When invoked 337 The same file — it is already loaded in full.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5 $0.00337 $0.00337
Opus 5 $0.00169 $0.00169
Sonnet 5 $0.00067 $0.00067
Haiku 4.5 $0.00034 $0.00034

Measured 5d ago against content hash 878356b5a10f, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

mcp-feedback scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 5d ago.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

.cursor/rules/mcp-feedback.mdc · 31 lines

What it actually says

Interactive Feedback MCP Rules

Core Rules

  1. If a request or instruction is unclear, use the interactive_feedback tool to ask the user clarifying questions before proceeding. Do not make assumptions.
  2. Provide predefined options via the interactive_feedback MCP tool whenever possible to facilitate quick decision-making.
  3. Each time you are about to complete a user request, call the interactive_feedback tool to ask for user feedback before finalizing. If the feedback is empty, you may end the request and must not call the tool in a loop.

Retry on Intermittent Failures

Cursor has a known bug where CallMcpTool occasionally reports "Tool not found" even though the MCP server is running. If this happens:

  1. Retry the interactive_feedback call up to 2 times before giving up.
  2. If all retries fail, fall back to the built-in AskQuestion tool.

Fallback

If the interactive_feedback tool call fails (timeout, connection error, or UI launch failure), use the built-in AskQuestion tool as a fallback.

Timeout Handling

  • The feedback window may remain open for extended periods while the user reviews or provides detailed feedback.
  • If the tool returns a message containing [心跳], it means the soft timeout was reached. Re-invoke interactive_feedback to continue the conversation.
  • Do not assume a timeout means the user is unresponsive — they may simply need more time.
Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. 5d ago First seen · 31 lines · 337 tokens per session scan A 878356b5a10f

Subscribe to this mod's changes

mcp-feedback is a cursor rule published in the GitHub repository dragonstylecc/Interactive-Feedback-With-Capture-MCP (4 stars, last pushed 5mo ago), licensed MIT. It adds 337 tokens to every session, about $0.0017 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.