Document /feedback command in CLI reference and data usage docs

Resolved 💬 4 comments Opened Feb 25, 2026 by ert485 Closed Mar 26, 2026

What's Missing?

The /feedback command exists in Claude Code (described as "Submit feedback about Claude Code") but is not documented anywhere:

  • The CLI reference lists slash commands but does not mention /feedback
  • The data usage docs describe what /bug sends but make no mention of /feedback

What Should Be Documented?

  1. What data does /feedback send? Does it include the full conversation transcript like /bug? Or just the feedback text?
  2. How does /feedback differ from /bug? Are they the same command under different names, or do they have different behavior/data collection?
  3. What is the retention policy for data submitted via /feedback?
  4. How to opt out of /feedback (is it the same DISABLE_BUG_COMMAND=1 env var?)

Users need this information to make informed decisions about submitting feedback, especially when working with sensitive code that may be in the conversation transcript. (Now I understand that there are 3 steps that confirm what will be sent, but it isn't clear before running the command that it isn't sent immediately -- I would look at the docs to determine that)

Environment

  • Claude Code version: 2.1.56
  • Platform: Anthropic API
  • OS: macOS (Darwin 25.2.0)
  • Shell: zsh

Related

  • #28563 — /feedback command fails silently on retry
  • #22971 — /feedback command exits prematurely

View original on GitHub ↗

This issue has 4 comments on GitHub. Read the full discussion on GitHub ↗