| name | evolution-report |
| description | 进化报告 — 分析系统运行趋势，识别改进机会，生成周度进化建议 |
| version | 1.0.0 |
| license | MIT |
| metadata | {"author":"blackplume233","actant-tags":"evolution,analytics,improvement,maintainer","layer":"kernel","agent":"maintainer"} |
Evolution Report
Overview
Analyze system operation trends, identify improvement opportunities, and generate weekly evolution recommendations.
Trend Analysis
- Performance: Latency percentiles, throughput over time
- Stability: Error rate, crash frequency, recovery time
- Usage: Agent invocations, skill usage, template adoption
- Resource: Disk growth, memory patterns, temp file accumulation
Performance Benchmarks
- Compare current week vs prior week for key metrics
- Flag regressions: "Latency p99 increased 2x vs last week"
- Highlight improvements: "Error rate down 30% after config fix"
Error Pattern Aggregation
- Group errors by type, Agent, and time window
- Identify top N recurring issues
- Correlate with deployments or config changes
Improvement Proposals
- Quick wins: Low-effort fixes with high impact
- Technical debt: Areas needing refactor or cleanup
- Capacity: Scaling or resource limits to consider
- Skills: Underused or overused skills; suggest additions/removals
Skill Effectiveness Scoring
- Track: skill invocation count, success rate, user satisfaction (if available)
- Recommend: deprecate low-value skills, promote high-value ones
- Report: "Skill X used 50x, 95% success; Skill Y used 2x, 30% success — consider deprecating Y"
