From claude-code-dev
Reviews feedback captured from Claude Code sessions, processes sessions with LLM analysis, reviews pending items, and exports to GitHub issues for Product Forge.
npx claudepluginhub jpoutrin/product-forge --plugin claude-code-devThis skill uses the workspace's default tool permissions.
Review feedback captured from your Claude Code sessions and sync to Product Forge.
Analyzes conversation for skill feedback and creates labeled GitHub issues for bugs, enhancements, and positive outcomes. Use when skills fail, suggest better patterns, or perform well.
Analyzes Claude Code sessions to improve CLAUDE.md instructions and capture learnings. Quick mode suggests CLAUDE.md tweaks; deep mode reviews problems, patterns, preferences, and gaps.
Submits bug reports, feature requests, SDLC patterns, and workflow improvements to Claude SDLC wizard GitHub repo via gh issue create. Privacy-first with explicit permission before scanning setup files.
Share bugs, ideas, or general feedback.
Review feedback captured from your Claude Code sessions and sync to Product Forge.
/sync-feedback # Show status (default)
/sync-feedback --process # Process captured sessions with LLM analysis
/sync-feedback --review # Review pending feedback
/sync-feedback --export # Export to GitHub issues format
| Option | Description |
|---|---|
--status | Show feedback statistics across all projects |
--process | Process captured sessions with LLM to extract feedback |
--review | Interactive review of pending feedback items |
--export | Export reviewed feedback to GitHub issue format |
All feedback is stored in ~/.claude/learnings/:
~/.claude/learnings/
├── projects.json # Registry of opted-in projects
├── stats.json # Global statistics
└── projects/
└── {project-slug}/
└── feedback/
├── improvement/
├── skill-idea/
├── command-idea/
├── bug-report/
└── pattern/
1. Enable hooks → /enable-feedback-hooks
2. Work normally → Sessions captured on exit (fast, no LLM)
3. Process → /sync-feedback --process (LLM analyzes sessions)
4. Review → /sync-feedback --review
5. Submit → /sync-feedback --export → Create GitHub issue
When the user runs this command:
Run the process-sessions script:
python3 ${CLAUDE_PLUGIN_ROOT}/scripts/hooks/process-sessions.py --all
The script will:
~/.claude/learnings/sessions/claude -p to analyze messages for feedbackDisplay results:
Processing sessions...
[process-sessions] Processing session abc12345 for project 'product-forge'
[process-sessions] Saved: improvement-add-validation.md
[process-sessions] Saved: skill-idea-log-checker.md
Processed 3 sessions, found 5 feedback items
Run /sync-feedback --review to review extracted feedback.
Read statistics from ~/.claude/learnings/stats.json
Read projects from ~/.claude/learnings/projects.json
Count pending items by scanning ~/.claude/learnings/projects/*/feedback/**/*.md
Display summary:
Product Forge Feedback Status
==============================
Total feedback captured: 15
Pending review: 8
By Type:
improvement: 6 (4 pending)
skill-idea: 4 (2 pending)
command-idea: 2 (1 pending)
bug-report: 2 (1 pending)
pattern: 1 (0 pending)
By Project:
product-forge: 7 items
my-django-app: 5 items
frontend-app: 3 items
Run /sync-feedback --review to review pending items.
Find all pending feedback files:
find ~/.claude/learnings/projects -name "*.md" -path "*/feedback/*"
For each pending item, display:
[1/8] improvement - product-forge
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
Title: Add status filtering to /task-list command
Description:
The /task-list command shows all tasks regardless of status.
Adding --status flag would help filter to pending/completed items.
Target: plugins/product-design/commands/task-list.md
Actions:
[a] Approve - Mark as reviewed, ready for sync
[d] Dismiss - Remove this feedback
[s] Skip - Review later
[e] Edit - Open in editor
[q] Quit - Exit review
Select action:
Process user action:
reviewed~/.claude/learnings/synced/dismissed/Show summary at end:
Review complete:
Approved: 5
Dismissed: 2
Skipped: 1
Run /sync-feedback --export to generate GitHub issues.
Find reviewed feedback (status: reviewed):
grep -l "status: reviewed" ~/.claude/learnings/projects/*/feedback/**/*.md
For each reviewed item, generate GitHub issue format:
## Issue: [improvement] Add status filtering to /task-list
**Type**: improvement
**Target**: plugins/product-design/commands/task-list.md
**Source Project**: product-forge
### Description
The /task-list command shows all tasks regardless of status.
Adding --status flag would help filter to pending/completed items.
### Suggested Implementation
[From feedback description]
---
Captured via Product Forge feedback hooks
Offer submission options:
5 items ready for export:
Options:
[1] Copy to clipboard (one at a time)
[2] Create GitHub issue via gh CLI
[3] Export to file (feedback-export.md)
[4] Cancel
Select option:
After export, move items to ~/.claude/learnings/synced/
Each feedback file uses YAML frontmatter:
---
type: improvement
status: pending
captured: 2026-01-06T15:30:00Z
session_id: abc123
project: my-django-app
repo: https://github.com/user/my-django-app
target: plugins/python-experts/skills/django-api
---
# Add pagination guidance to django-api skill
The django-api skill covers endpoint creation but lacks
guidance on pagination patterns for large datasets.
pending → reviewed → synced
↓
dismissed
--status regularly to see accumulated feedback