Back to Claude Mem

Phase 01: Close Duplicates & Stale Issues

.maestro/playbooks/2026-03-29-Issues-Triage-3-29-2026/Phase-01-Close-Duplicates-And-Stale-Issues.md

13.15.38.5 KB
Original Source

Phase 01: Close Duplicates & Stale Issues

This phase performs the immediate triage cleanup of the claude-mem GitHub issue backlog (thedotmack/claude-mem). By closing ~17 duplicate, vague, and outdated issues with clear explanatory comments, and labeling all remaining issues by root-cause group, we reduce noise and create a clean, organized backlog for the code-fix phases that follow. This phase requires no code changes — only gh CLI operations against the GitHub API.

Tasks

  • Create GitHub labels for root-cause categorization on thedotmack/claude-mem. ✓ All 10 labels created: root:worker-lifecycle, root:windows, root:chromadb, root:project-scoping, root:session-integrity, root:security, root:installer, root:mcp-schema, root:aws-bedrock, triage:closable Use gh label create for each (skip if already exists — use gh label list first to check):

    • root:worker-lifecycle (color: #d73a4a) — Worker startup, shutdown, zombie processes
    • root:windows (color: #0075ca) — Windows-specific platform bugs
    • root:chromadb (color: #7057ff) — ChromaDB/MCP subprocess issues
    • root:project-scoping (color: #008672) — Project identity, isolation, namespace collisions
    • root:session-integrity (color: #e4e669) — Session data, pending queue, finalization
    • root:security (color: #b60205) — Security vulnerabilities
    • root:installer (color: #fbca04) — Setup, install, marketplace path issues
    • root:mcp-schema (color: #1d76db) — MCP tool registration, schema issues
    • root:aws-bedrock (color: #5319e7) — AWS Bedrock integration
    • triage:closable (color: #cccccc) — Marked for closure during triage
  • Close the 6 duplicate __dirname issues, keeping #1410 as the canonical issue. For each of #1419, #1428, #1433, #1434, #1437, #1438: run gh issue close <NUMBER> -R thedotmack/claude-mem -c "Closing as duplicate of #1410. All 7 reports trace to the same root cause: hardcoded __dirname in the esbuild bundle output (worker-service.cjs). Fix will be tracked in #1410.". Then add label root:worker-lifecycle to #1410. ✓ All 6 duplicates (#1419, #1428, #1433, #1434, #1437, #1438) closed with explanatory comments. Label root:worker-lifecycle added to #1410.

  • Close the duplicate setup.sh pair and duplicate security pair: ✓ #1396 closed as dup of #1340 (root:installer label added), #1521 closed as dup of #1285 (root:security label added)

    • Close #1396 as duplicate of #1340: gh issue close 1396 -R thedotmack/claude-mem -c "Duplicate of #1340 — both report the same missing scripts/setup.sh issue." Add label root:installer to #1340.
    • Close #1521 as duplicate of #1285: gh issue close 1521 -R thedotmack/claude-mem -c "Duplicate of #1285 — both report command injection in GitHub Actions workflow." Add label root:security to #1285.
  • Close vague, no-info, and non-actionable issues. For each, use gh issue close with a polite explanatory comment: ✓ All 7 vague/non-actionable issues closed (#1492, #1436, #1362, #1385, #1488, #1459, #1378) with explanatory comments.

    • #1492 ("Bug Report") — "Closing: this issue has only a generic title with no reproduction steps, error logs, or version info. Please reopen with specific details if the issue persists."
    • #1436 ("bug") — "Closing: no description, steps to reproduce, or error output provided. Please reopen with details if still relevant."
    • #1362 (Chinese, no repro) — "Closing: unable to action without reproduction steps or error details. If you can provide steps to reproduce (in any language), please reopen."
    • #1385 ("mem still not save") — "Closing: insufficient details to diagnose. If you can share your OS, claude-mem version, and error logs, please reopen a new issue."
    • #1488 ("Consumed 25% of my tokens") — "Closing: this appears to be a usage concern rather than a bug. Token consumption depends on context window size and session length. If you believe there is a specific bug causing excessive token usage, please open a new issue with reproduction steps."
    • #1459 ("Featured in awesome-claude-code-workflows") — "Closing: this is a community celebration rather than an actionable issue. Thank you for the recognition! 🎉"
    • #1378 ("Test failure: test run failed") — "Closing: no test output, stack trace, or version info provided. This may have been fixed in later releases. Please reopen with details if still occurring."
  • Close likely-fixed and superseded issues: ✓ All 3 closed (#1268 regression fixed in 10.6.x+, #1137 superseded by #1262, #1219 stale version bump).

    • #1268 — gh issue close 1268 -R thedotmack/claude-mem -c "This regression appears to have been fixed in subsequent releases (10.6.x+). If you're still experiencing this on the latest version, please reopen with your version number and error details."
    • #1137 — gh issue close 1137 -R thedotmack/claude-mem -c "Superseded by #1262 which covers the broader plan mode pending messages issue. Closing to consolidate tracking."
    • #1219 (version bump to 10.4.1) — gh issue close 1219 -R thedotmack/claude-mem -c "This version has long since been superseded. The current version is well past 10.4.1. Closing as stale."
  • Label all remaining open issues by root-cause group. ✓ 76 label applications across 9 root-cause groups on 93 open issues. Multi-label issues: #1489 (windows+chromadb), #1225 (windows+chromadb), #1234 (worker-lifecycle+session-integrity). Breakdown: root:worker-lifecycle (21), root:windows (13), root:chromadb (9), root:project-scoping (6), root:session-integrity (13), root:security (4), root:installer (7), root:mcp-schema (4), root:aws-bedrock (2). ~16 feature requests/misc kept without root labels. First run gh issue list -R thedotmack/claude-mem --state open --limit 200 --json number,title,labels to get the current open list, then apply labels using gh issue edit <NUMBER> -R thedotmack/claude-mem --add-label <LABEL>. Group assignments:

    • root:worker-lifecycle: #1410, and any issues mentioning worker startup failures, cold start, zombie processes, version mismatch, daemon spawn, "worker not running", "ECONNREFUSED", port binding, health check failures, or ensureWorkerRunning
    • root:windows: Any issues mentioning Windows, PowerShell, CRLF, win32, zombie sockets on Windows, .exe, taskkill, or Windows Terminal
    • root:chromadb: Any issues mentioning ChromaDB, chroma-mcp, uvx, vector search failures, CPU spin from chroma, or embedding errors
    • root:project-scoping: Any issues mentioning project name collisions, basename, monorepo isolation, workspace bleed, or data from wrong project
    • root:session-integrity: Any issues mentioning early finalization, duplicate observations, empty summaries, pending queue growth, or lost data
    • root:security: #1285, #1204, and any issues mentioning command injection, file write vulnerability, port collision data leakage, or permission bypass
    • root:installer: #1340, and any issues mentioning setup, install, marketplace path, allowlist, plugin.json, or MCP registration failures
    • root:mcp-schema: Any issues mentioning empty inputSchema, MCP tool ID mismatch, or SSE broadcast
    • root:aws-bedrock: Any issues mentioning Bedrock, AWS SDK, or Bedrock environment variables
    • Issues fitting multiple categories should receive multiple labels. Feature requests should keep their existing enhancement label if present.
  • Generate a structured triage summary report. ✓ Report written to Working/triage-report.md. 112 issues before triage, 18 closed (8 duplicate, 7 vague, 3 stale), 94 remaining. 15 P0, 54 P1, 25 P2. All 94 open issues categorized by 9 root-cause labels plus 15 unlabeled standalone items. Create a markdown file at /Users/alexnewman/Scripts/claude-mem/Auto Run Docs/2026-03-29-Issues-Triage-3-29-2026/Working/triage-report.md with YAML front matter:

    yaml
    ---
    type: report
    title: "Issues Triage Report — 2026-03-29"
    created: 2026-03-29
    tags:
      - triage
      - issues
      - claude-mem
    ---
    

    The report should contain:

    • Total issues before triage vs after (with count of closed)
    • Table of closed issues with number, title, and reason (duplicate/stale/vague)
    • Table of remaining open issues grouped by root-cause label, with issue number, title, and priority (P0/P1/P2)
    • Priority criteria: P0 = blocks all users or security vulnerability, P1 = affects large user segment, P2 = nice-to-have or single-user report
    • A "Next Steps" section listing the 6 remaining fix phases with brief descriptions