[Feature Request] Local Archive of /feedback Submissions

Bug Description
Feature Request: Local Archive of /feedback Submissions

From: Jordan Dea-Mattson
GitHub: @jordan-of (OrdinaryFolk), @jordandm (personal)
Email: jordan-of@users.noreply.github.com

Problem

When /feedback is submitted, the feedback text is sent to Anthropic and optionally a GitHub issue is opened. But no local copy is kept. The user has no record of:

• What they reported
• When they reported it
• The feedback item ID Anthropic generates
• The GitHub issue URL (if one was opened)

If the conversation transcript is later compressed or the session ends, the feedback text is lost from the user's perspective. The only way to recover it is to find the GitHub issue (if one was
created) or hope Anthropic's internal system retained the exact text.

Requested Behavior

After each /feedback submission, save a local record. The archive location should be configurable at two levels:

• User-level: ~/.claude/feedback/ — default, always available, spans all projects
• Project-level: <project>/.claude/feedback/ — optional, keeps feedback visible in the repo context where it originated
• Both: configurable via ~/.claude/settings.json or project settings.json (e.g., "feedback_archive": "both" / "user" / "project")

When set to "both", the same file is written to both locations. Project-level archives can be gitignored or committed depending on team preference.

Each file should contain:

• Timestamp of submission
• Feedback item ID returned by Anthropic's system
• GitHub issue URL (if one was opened, or "not opened")
• Session ID and project context
• The full feedback text as submitted

Example:

---
date: 2026-03-21T02:35:20Z
feedback_id: fb_abc123
github_issue: https://github.com/anthropics/claude-code/issues/36909
session: 0ac70dee-ed27-41a5-b3f6-569be4c8d34b
project: /Users/jordan_of/code/monofolk
---

## Session naming and machine-readable state indicators

(full feedback text here)

Why This Matters

• Accountability — users should have a record of what they reported and when
• Follow-up — easy to check status of past reports without searching GitHub
• Deduplication — before filing new feedback, check what you've already reported
• Loss prevention — feedback drafted collaboratively with Claude during a session disappears when the session ends unless manually saved

This is a small change (write one file after submission) with outsized value for active reporters.

Environment Info

• Platform: darwin
• Terminal: iTerm.app
• Version: 2.1.80
• Feedback ID: b846ee2f-6bf5-4106-b206-16da41301d76

Errors

[{"error":"Error: NON-FATAL: Lock acquisition failed for /Users/jordan_of/.local/share/claude/versions/2.1.80 (expected in multi-process scenarios)\n    at YST (/$bunfs/root/src/entrypoints/cli.js:2903:2174)\n    at fBq (/$bunfs/root/src/entrypoints/cli.js:2903:1318)\n    at processTicksAndRejections (native:7:39)","timestamp":"2026-03-20T06:22:33.390Z"},{"error":"MaxFileReadTokenExceededError: File content (10820 tokens) exceeds maximum allowed tokens (10000). Use offset and limit parameters to read specific portions of the file, or search for specific content instead of reading the whole file.\n    at z_9 (/$bunfs/root/src/entrypoints/cli.js:4268:23437)\n    at processTicksAndRejections (native:7:39)","timestamp":"2026-03-20T06:36:03.196Z"},{"error":"MaxFileReadTokenExceededError: File content (17524 tokens) exceeds maximum allowed tokens (10000). Use offset and limit parameters to read specific portions of the file, or search for specific content instead of reading the whole file.\n    at z_9 (/$bunfs/root/src/entrypoints/cli.js:4268:23437)\n    at processTicksAndRejections (native:7:39)","timestamp":"2026-03-20T06:39:17.705Z"},{"error":"MaxFileReadTokenExceededError: File content (14803 tokens) exceeds maximum allowed tokens (10000). Use offset and limit parameters to read specific portions of the file, or search for specific content instead of reading the whole file.\n    at z_9 (/$bunfs/root/src/entrypoints/cli.js:4268:23437)\n    at processTicksAndRejections (native:7:39)","timestamp":"2026-03-20T06:39:20.137Z"},{"error":"MaxFileReadTokenExceededError: File content (20579 tokens) exceeds maximum allowed tokens (10000). Use offset and limit parameters to read specific portions of the file, or search for specific content instead of reading the whole file.\n    at z_9 (/$bunfs/root/src/entrypoints/cli.js:4268:23437)\n    at processTicksAndRejections (native:7:39)","timestamp":"2026-03-20T06:58:34.461Z"},{"error":"MaxFileReadTokenExceededError: File content (11735 tokens) exceeds maximum allowed tokens (10000). Use offset and limit parameters to read specific portions of the file, or search for specific content instead of reading the whole file.\n    at z_9 (/$bunfs/root/src/entrypoints/cli.js:4268:23437)\n    at processTicksAndRejections (native:7:39)","timestamp":"2026-03-20…

Note: Content was truncated.