agent-orchestration-improve-agent
Systematic improvement of existing agents through performance analysis, prompt engineering, and continuous iteration.
用 Codex 或 Claude 帮你安装 复制这段 Prompt,粘贴到 Codex、Claude 或其他助手里,让它检查 Skill 页面并帮你完成安装。
菜单
Systematic improvement of existing agents through performance analysis, prompt engineering, and continuous iteration.
用 Codex 或 Claude 帮你安装 复制这段 Prompt,粘贴到 Codex、Claude 或其他助手里,让它检查 Skill 页面并帮你完成安装。
基于 SOC 职业分类
Build production-ready LLM applications, RAG systems, and intelligent agents. Use when implementing AI features, chatbots, vector search, or agent orchestration.
Write, rewrite, or improve marketing copy for any page type. Use when crafting homepage, landing page, pricing, or feature page copy with conversion best practices.
Diagnose and audit SEO issues affecting crawlability, indexation, and rankings. Use when performing technical SEO reviews, meta tag audits, or organic performance health checks.
TypeScript and JavaScript expert for type-level programming, performance optimization, and modern tooling. Use when solving complex type issues, build problems, or migration decisions.
Design and implement durable workflow automation with Temporal, Inngest, and n8n. Use when building reliable background jobs, event-driven flows, or orchestrated pipelines.
Research a topic from the last 30 days on Reddit + X + Web, become an expert, and write copy-paste-ready prompts for the user's target tool.
| name | agent-orchestration-improve-agent |
| description | Systematic improvement of existing agents through performance analysis, prompt engineering, and continuous iteration. |
| risk | unknown |
| source | rootcastle-rei |
Systematic improvement of existing agents through performance analysis, prompt engineering, and continuous iteration.
[Extended thinking: Agent optimization requires a data-driven approach combining performance metrics, user feedback analysis, and advanced prompt engineering techniques. Success depends on systematic evaluation, targeted improvements, and rigorous testing with rollback capabilities for production safety.]
Comprehensive analysis of agent performance using context-manager for historical data collection.
Use: context-manager
Command: analyze-agent-performance $ARGUMENTS --days 30
Collect metrics including:
Identify recurring patterns in user interactions:
Categorize failures by root cause:
Generate quantitative baseline metrics:
Performance Baseline:
- Task Success Rate: [X%]
- Average Corrections per Task: [Y]
- Tool Call Efficiency: [Z%]
- User Satisfaction Score: [1-10]
- Average Response Latency: [Xms]
- Token Efficiency Ratio: [X:Y]
Apply advanced prompt optimization techniques using prompt-engineer agent.
Implement structured reasoning patterns:
Use: prompt-engineer
Technique: chain-of-thought-optimization
Curate high-quality examples from successful interactions:
Example structure:
Good Example:
Input: [User request]
Reasoning: [Step-by-step thought process]
Output: [Successful response]
Why this works: [Key success factors]
Bad Example:
Input: [Similar request]
Output: [Failed response]
Why this fails: [Specific issues]
Correct approach: [Fixed version]
Strengthen agent identity and capabilities:
Implement self-correction mechanisms:
Constitutional Principles:
1. Verify factual accuracy before responding
2. Self-check for potential biases or harmful content
3. Validate output format matches requirements
4. Ensure response completeness
5. Maintain consistency with previous responses
Add critique-and-revise loops:
Optimize response structure:
Comprehensive testing framework with A/B comparison.
Create representative test scenarios:
Test Categories:
1. Golden path scenarios (common successful cases)
2. Previously failed tasks (regression testing)
3. Edge cases and corner scenarios
4. Stress tests (complex, multi-step tasks)
5. Adversarial inputs (potential breaking points)
6. Cross-domain tasks (combining capabilities)
Compare original vs improved agent:
Use: parallel-test-runner
Config:
- Agent A: Original version
- Agent B: Improved version
- Test set: 100 representative tasks
- Metrics: Success rate, speed, token usage
- Evaluation: Blind human review + automated scoring
Statistical significance testing:
Comprehensive scoring framework:
Task-Level Metrics:
Quality Metrics:
Performance Metrics:
Structured human review process:
Safe rollout with monitoring and rollback capabilities.
Systematic versioning strategy:
Version Format: agent-name-v[MAJOR].[MINOR].[PATCH]
Example: customer-support-v2.3.1
MAJOR: Significant capability changes
MINOR: Prompt improvements, new examples
PATCH: Bug fixes, minor adjustments
Maintain version history:
Progressive deployment strategy:
Quick recovery mechanism:
Rollback Triggers:
- Success rate drops >10% from baseline
- Critical errors increase >5%
- User complaints spike
- Cost per task increases >20%
- Safety violations detected
Rollback Process:
1. Detect issue via monitoring
2. Alert team immediately
3. Switch to previous stable version
4. Analyze root cause
5. Fix and re-test before retry
Real-time performance tracking:
Agent improvement is successful when:
After 30 days of production use:
Establish regular improvement cadence:
Remember: Agent optimization is an iterative process. Each cycle builds upon previous learnings, gradually improving performance while maintaining stability and safety.
🏰 Rei Skills — Curated by Rootcastle Engineering & Innovation | Batuhan Ayrıbaş
Engineering Beyond Boundaries | admin@rootcastle.com