anthropics / anthropics/claude-code
[FEATURE] Disclose transcript scope and retention in the /bug and /feedback dialog before submission
- Dominant language
- Python
- Stars
- 145k
- Forks
- 23.1k
- PR merge metrics
- PR metrics pending
Description
### Preflight Checklist
- [x] I have searched [existing requests](https://github.com/anthropics/claude-code/issues?q=is%3Aissue%20label%3Aenhancement) and this feature hasn't been requested yet
- [x] This is a single feature request (not multiple features)
### Problem Statement
The /bug and /feedback dialog does not state at the point of submission what
is actually sent. Per https://code.claude.com/docs/en/data-usage, a report
carries a copy of the full conversation history for the selected session(s) —
code, file contents, and tool outputs — retained for 5 years.
None of that is legible in the dialog itself. The scope selector (current
session / 24h / 7d) implies some history is involved, but never says the
transcript itself is included, nor mentions the retention period. A user
reporting a minor UI issue (in my case, an ambiguous label) has no in-dialog
indication that their report discloses the entire session. If that session
contains proprietary code or security-sensitive material, the user submits it
without realizing. I discovered what I had sent only afterward, by reading the
data-usage docs.
Had the consequence been visible at the decision point, I would have filed the
same report from a clean session — the workaround is trivial once you know.
Related: #93491, #72156, #74814 — those ask for opt-out, redaction, and safer consent mechanics respectively; this asks only that the existing dialog state what is sent and for how long it's retained, matching the disclosure pattern the new Claude-drafted feedback cards already use
### Proposed Solution
One or two sentences in the dialog, before the submit action, e.g.:
"This sends the full transcript of the selected session(s), including code
and tool outputs, to Anthropic. Retained 5 years; not used to train
generative models."
Optionally a "Learn more" link to the data-usage page. No behavior change
needed — just making the existing policy visible where the decision is made.
### Alternative Solutions
_No response_
### Priority
Not blocking work but creates a privacy concern/issue
### Feature Category
CLI commands and flags
### Use Case Example
_No response_
### Additional Context
_No response_
Contributor guide
No contributing guide indexed for this repository
Research direction
Locate the implementation of the /bug and /feedback dialog, then compare its submission flow with the data-usage documentation linked in the issue. Done means the dialog clearly states that selected session transcripts, including code and tool outputs, are sent and retained for five years before submission; an optional Learn more link is also proposed.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- python
- Domain
- cli, security
- Issue type
- Feature
- Difficulty
- 3/5
- Estimated time
- 1-2 days
- Activity status
- Active
- Clarity
- Mostly clear
- Newbie friendliness
- 58/100