{
  "markdown": "# find-conversation\n\nA Claude Code plugin to instantly search and resume past conversations by keyword.\n\n## What it does\n\nSearches all your Claude Code conversation history across every project and returns a formatted table with:\n\n- Session slug and match count\n- Project path (decoded from Claude's internal format)\n- Start and last-active timestamps\n- Last user message preview\n- First matching snippet for context\n- Ready-to-copy `claude --resume` commands\n\n## Install\n\n```bash\n/plugin marketplace add akzarma/claude-find-conversation\n/plugin install find-conversation@akzarma-tools\n```\n\n## Usage\n\n```bash\n/find-conversation <search terms>\n```\n\n### Examples\n\n```bash\n/find-conversation deployment vercel\n/find-conversation auth middleware bug\n/find-conversation database migration\n```\n\n## How it works\n\n1. **Binary pre-filter** — scans conversation JSONL files in 1MB chunks to skip non-matching files instantly\n2. **Single-pass parsing** — reads each matching file once, extracting metadata, match counts, and message previews\n3. **Efficient last-line seek** — reads timestamps from file ends without loading entire files into memory\n4. **Smart path decoding** — reconstructs real filesystem paths from Claude's dash-encoded project folder names using greedy filesystem matching\n\nNo external dependencies. Python 3.8+ only (ships with macOS and Linux).\n\n## Requirements\n\n- Claude Code (macOS or Linux)\n- Python 3.8+ (pre-installed on all supported platforms)\n\n## License\n\nMIT\n",
  "bytes": 1480,
  "sha": "2b9028b68c3bbf1a152320fa2cb824740e7234ac87cb1c50585efb321b931865",
  "repo_slug": "akzarma/claude-find-conversation",
  "fonte": "repo",
  "truncated": false,
  "api": "https://agentalog.com/api/listings/plg_akzarma_claude_find_conversation_find_co_1f2c65fc/readme"
}