analyze-context-routerlisted

# Memory Retrieval Analysis > **Requires the `multiplai-context` plugin.** This skill analyzes that plugin's runtime artifacts — the `context-router-*.log` retrieval logs, `context-router.py` hook, and `memory-catalog.json`. Without multiplai-context installed and active, there are no logs to analyze. Analyze the memory retrieval log files to assess routing quality, identify failure modes, and track improvement over time. Uses a watermark file to partition pre/post-fix data so metrics reflect only the current system behavior. ## Key Files | File | Purpose | |------|---------| | `$CLAUDE_CONFIG_DIR/context-router-*.log` | Daily retrieval logs (7-day retention) | | `$CLAUDE_CONFIG_DIR/context-router-watermark.json` | Analysis watermark: last analysis timestamp, baseline metrics, deployment info | | `$CLAUDE_CONFIG_DIR/hooks/context-router.py` | The retrieval hook source (for recommending code changes) | | `$CLAUDE_CONFIG_DIR/memory/memory-catalog.json` | File catalog used for routing (for recommending catalog changes) | ## Workflow ### Step 1: Read Watermark Read `$CLAUDE_CONFIG_DIR/context-router-watermark.json`. Extract: - `deployed_at` — timestamp of last improvement deployment - `last_analysis` — timestamp of last analysis run - `baseline_metrics` — metrics from the previous analysis (for delta comparison) If no watermark exists, this is the first run. Analyze all available logs and create the watermark at the end. ### Step 2: Discover and Filter Logs Glob `$CLAUD