Skill v1.0.1
currentAutomated scan100/100+3 new
version: "1.0.1" name: mine description: 'Extract a specific topic from a single session JSONL file using 4 parallel subagents. Use when user says "mine", "extract topic", "find in session", or wants to search a session for a keyword.' argument-hint: "<keyword> [session-id]" zombie: true origin: arra-symbiosis-skills metadata: internal: true
/mine — Oracle Mining
Extract a specific topic from a single session JSONL file.
Usage
/mine worktree # Mine latest session for "worktree"/mine ghost-agents abc12345 # Mine specific session
How It Works
Detects whether the first argument is a session ID (hex) or treats it as a keyword for the most recent session. Launches 4 parallel subagents that all read the same JSONL file and filter by keyword (case-insensitive):
| Agent | Role | |
|---|---|---|
| What You Said | Human messages containing keyword, timeline of mentions | |
| What AI Did | Assistant responses, tool calls, bash commands with keyword | |
| Files + Code | Write/Edit operations, git commits, actual code snippets (truncated 50 lines) | |
| Connections | Spawned agents, skill invocations, MCP calls, co-occurring themes, URLs |
Output
Compiled report with timestamps, keyword frequency, and sections for each agent's findings. Concludes with 1-3 sentence summary.
Rules
- Strictly processes one JSONL file
- Uses local timezone conversion
- Case-insensitive matching
- Displays actual code rather than summaries
- No output writes to trace logs
ARGUMENTS: $ARGUMENTS