evolve-report

name	evolve-report
description	View evolution daily reports and usage statistics. Shows session data, tool usage, token consumption, and trend analysis. Trigger: /evolve-report, evolve-report, 查看报告, 进化报告, /evolve-stats, evolve-stats.

Evolve-Report — Report & Statistics Viewer

Purpose

Display daily analysis reports and usage statistics from collected data.

When Triggered

User says "/evolve-report", "evolve-report", "查看报告", "进化报告", "/evolve-stats", "evolve-stats", or wants to view evolution data.

Data Paths

• Daily reports: ~/.pi/agent/evolution-data/daily-reports/*.json (Only .json files are Python analyzer output. .md files are legacy from old evolution-engine, ignore them.)
• Daily summaries: ~/.pi/agent/evolution-data/daily/*.json (Generated by usage-tracker extension.)
• Metrics history: ~/.pi/agent/evolution-data/metrics-history.json

Procedure

Parse Command

• No args → show today's report
• Date string (YYYY-MM-DD) → show that date's report
• --list → list all available reports
• --stats or "/evolve-stats" → show usage statistics overview

Show Report

1. Use bash tool to check available reports:

   ls ~/.pi/agent/evolution-data/daily-reports/*.json 2>/dev/null
   

2. Read the requested report using read tool

3. Present key information in a structured format:

   • Session count and duration
   • Tool call statistics (most used, error rates)
   • Token consumption (input/output)
   • Anomalies and signals
   • Improvement suggestions (if any in the report)

4. New Dimensions — if present in the report, display additional sections:

   Tool Parameter Errors (tool_error_stats)

   Display when tool_error_stats field exists:

   • Param error rate per tool (highlight tools with rate > 25%)
   • Top parameter error patterns with occurrence counts
   • Actionable issues from feedback-records (if any)

   Goal Task Quality (goal_quality_stats)

   Display when goal_quality_stats field exists:

   • Task completion rate (highlight if < 50%)
   • Average tasks per goal
   • Evidence quality distribution (high/medium/low)
   • Actionable issues: task splitting, evidence requirements

   Subagent Efficiency (subagent_stats)

   Display when subagent_stats field exists:

   • Success / failure / retry counts
   • Failure rate (highlight if > 20%)
   • Retry rate (highlight if > 15%)
   • Actionable issues: task prompt optimization, task splitting

   Compact & Context Efficiency (compact_stats + context_stats)

   Display when either field exists:

   • Average compacts per session (highlight if ≥ 3)
   • Context utilization trend (rising / stable / falling)
   • Actionable issues: context management, tool output optimization

   Workflow Efficiency (workflow_stats)

   Display when workflow_stats field exists:

   • Per-phase duration breakdown (highlight phase > 50% of total)
   • Gate retry frequency per phase
   • Actionable issues: workflow optimization, gate criteria tuning

   Todo Usage (todo_stats)

   Display when todo_stats field exists:

   • Completion / abandon / in-progress counts
   • Abandon rate (highlight if > 25%)
   • Actionable issues: todo usage pattern optimization

   Formatting guidelines for new dimensions:

   • Use a consistent section header format: ### Dimension Name
   • Show metrics as key-value pairs or simple tables
   • Prefix actionable issues with ⚠ marker
   • If a dimension field is missing from the report, skip that section entirely
   • Group all new dimension sections under a ## Extended Metrics heading

List Reports

ls -1 ~/.pi/agent/evolution-data/daily-reports/*.json 2>/dev/null | xargs -I{} basename {} .json

Present as a numbered list of available dates.

Show Statistics

1. Read multiple daily/*.json files for the requested period
2. Aggregate and present:
   • Total sessions, tool calls, tokens
   • Per-tool usage breakdown
   • Day-over-day trends
   • Top skills triggered
   • Error patterns

Use tables and summaries for readability. Highlight significant changes.