diegosouzapw/awesome-omni-skill / skills/ai-agents/agent-evaluation-cdalsoniii
Agent Evaluation
VerifiedEvaluate agent performance using a structured scoring rubric
Agent Evaluation Skill
Evaluate agent performance using a structured scoring rubric.
Trigger Conditions
- Agent configuration change
- Evaluation cadence (monthly)
- User invokes with "evaluate agents" or "agent scorecard"
Input Contract
- Required: Agent(s) to evaluate
- Required: Evaluation criteria or rubric
- Optional: Baseline scores from prior evaluation
Output Contract
- Evaluation scorecard (500-point rubric)
- Per-dimension scores and findings
- Improvement recommendations
- Comparison against baseline
Tool Permissions
- Read: Agent configs, agent output logs, telemetry
- Write: Evaluation reports
- Search: Agent invocation history
Execution Steps
- Load evaluation rubric (architecture 100, security 100, ops 100, testing 100, docs 100)
- For each agent, review recent outputs and effectiveness
- Score each dimension with evidence
- Compare against baseline scores
- Identify improvement opportunities
- Generate scorecard and recommendations
Success Criteria
- All dimensions scored with evidence
- Comparison against prior evaluation
- Top 3 improvement recommendations per agent
- Overall portfolio health assessment
Escalation Rules
- Escalate if any agent scores below 50% on any dimension
- Escalate if agent evaluation reveals conflicting outputs between agents
Example Invocations
Input: "Evaluate the security-specialist agent effectiveness"
Output: Scorecard: Security (85/100), Architecture (70/100), Ops (75/100), Testing (60/100), Docs (80/100). Total: 370/500. Findings: strong CVE detection but weak test coverage recommendations, documentation quality high but missing escalation follow-through. Top improvement: integrate with testing-specialist for security test gap analysis.
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| reverse-engineer-anythingmorluto/reaReverse engineer native, managed, Electron/JavaScript, packaged, firmware, and browser targets with REA, and build agent environments (AgentEnv bundles) of web applications from a live instance or a capture, and of Linux desktop apps with a task from a request. Use shipped-artifact or requested runtime evidence to explain features, compare versions, decompile code, or guide a reconstruction. Skip REA for ordinary source-repository architecture analysis. | Stars: 471k | Updated | Claude CodeCodex | Verified | Install |
| diagnosing-superpowersobra/superpowersUse when a superpowers session went wrong and your human partner wants to know why — repeated work, ignored plans, stumbles, poor results, a skill that didn't fire, "it took too long", "why is it so expensive", "what is it doing" — or wants to build a bug report for the superpowers maintainers, for the current session or a past one identified by id or path, on any harness. | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| dispatching-parallel-agentsobra/superpowersUse when facing 2+ independent tasks that can be worked on without shared state or sequential dependencies | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| executing-plansobra/superpowersUse when executing an implementation plan in the current session as the implementer yourself — your human partner chose inline execution, or no subagent tool is available | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |