consolidate-discovery

Consolidate Discovery

Discovers learnings that may need consolidation: duplicates and outdated content.

Usage

python3 consolidate-discovery.py [--mode all|duplicates|outdated] [--limit N]

Options

• --mode: Discovery type (default: all)
• --limit: Max items per category (default: 20)

Output Format

Compact JSON with file basenames, paths, and IDs:

{
  "status": "success",
  "total_documents": 187,
  "duplicates": [
    {
      "files": [
        {"id": "abc", "file": "jwt-auth.md", "path": "/full/path/jwt-auth.md"},
        {"id": "def", "file": "jwt-auth.md", "path": "/other/path/jwt-auth.md"}
      ],
      "count": 2
    }
  ],
  "outdated": [
    {"id": "ghi", "file": "temp-fix.md", "path": "/path/temp-fix.md", "markers": ["temporary"]}
  ],
  "summary": {
    "duplicate_clusters": 3,
    "outdated_candidates": 2,
    "limit_applied": 20
  }
}

Notes

• Uses tight similarity threshold (0.25) for true duplicates
• Output is compact to avoid token bloat
• Use consolidate-actions get to fetch full content when needed