evaluatelisted

# Evaluate - PDCA Evaluation & Improvement A skill that analyzes sales activity result data, reports on performance, and applies the improvements it still owns — numeric priorities and targeting / search keywords. Messaging (subject lines) and channel ranking are now optimized deterministically by the daily lever tick; evaluate **reads and narrates** those, it does not rewrite them as strategy prose. **Before starting:** `Read` `${CLAUDE_PLUGIN_ROOT}/references/workspace-conventions.md` and follow the cross-cutting conventions there (data storage, MCP error handling, document writes, output discipline). ## Steps ### 1. Data Collection - Project ID: `$0` (required) In parallel, call: - `mcp__plugin_leadace_api__get_eval_data` with `projectId: "$0"` - `mcp__plugin_leadace_api__get_rejection_feedback_summary` with `projectId: "$0"`, `windowDays: 30`, `scope: "tactical"` - `mcp__plugin_leadace_api__get_lever_state` with `projectId: "$0"` — current subject draw weights, channel affinity, and per-variant maturity (read-only) - `mcp__plugin_leadace_api__get_lever_decisions` with `projectId: "$0"` — the daily tick's decision history (newest first) for trend narration If `get_eval_data` returns a "Project not found" error, instruct the user to run `/leadace` first and **abort**. `get_eval_data` response includes: - `metrics`: totalOutreach, channelCounts, responseCounts, sentimentBreakdown, priorityResponseRate, statusCounts, channelResponseRate, variantResponseRate, inquiryOu