session-analyticslisted

## Routing Guidance - **CAPTURE**: 'capture session', 'collect artifacts', 'headless capture', 'run verification capture', 'profile a skill invocation', 'audit skill routing', 'measure hook in vivo', 'behavioral capture' - **INSPECT**: 'parse session', 'session metrics', 'session data', 'inspect transcript', 'session cost', 'what tools did I use', 'how long was this session', 'extract session data' - **REVIEW**: 'review session', 'what went wrong', 'behavioral review', 'anti-pattern', 'meta-review', 'session critique', 'why did it fail', 'investigate session' - **ISSUE**: 'create issue from review', 'file report', 'meta-issue', 'generate GitHub issue', 'make a bug report from session' - CONTRAST with diagnose: session-analytics analyzes session transcripts; diagnose analyzes code problems. - CONTRAST with code-review: session-analytics extracts behavioral patterns; code-review analyzes source code quality. ## What This Skill Changes **Default behavior:** Session data remains unstructured JSONL — tool calls, errors, and cost metrics require ad-hoc grep commands. Anti-patterns surface only when the user explicitly asks, and no systematic recording exists. Issue creation is manual and inconsistent. **With this skill:** Standardized three-mode extraction (INSPECT), behavioral diagnosis (REVIEW), and sanitized issue generation (ISSUE). The main agent delegates data extraction and analysis to subagents; it synthesizes results and manages the issue creation workflow. **Why mode