Agent skill for performance-benchmarker - invoke with $agent-performance-benchmarker
40
Quality
13%
Does it follow best practices?
Impact
81%
2.89xAverage score across 3 eval scenarios
Passed
No known issues
Optimize this skill with Tessl
npx tessl skill review --optimize ./.agents/skills/agent-performance-benchmarker/SKILL.mdThroughput measurement defaults and adaptive rate control
Default duration
0%
100%
Default initial rate
100%
100%
Default pattern and rampUp
0%
25%
Measurement interval
100%
100%
Rate increment value
0%
100%
Rate increase threshold
80%
100%
Rate decrease threshold
30%
100%
Optimal throughput definition
83%
100%
Sustainable throughput definition
0%
100%
Per-measurement latency fields
0%
100%
Throughput degradation flag
0%
0%
Without context: $0.7771 · 3m 7s · 30 turns · 36 in / 13,091 out tokens
With context: $0.8311 · 3m 8s · 27 turns · 319 in / 11,162 out tokens
Latency phase analysis and percentile reporting
Warmup performed
100%
100%
Default warmup size
0%
100%
Default sample size
0%
100%
Three latency phases
0%
100%
Total latency as sum of phases
0%
100%
Percentile set
66%
100%
Outlier identification
0%
100%
Phase contributionPercent
100%
100%
Per-phase percentile fields
50%
100%
Successful measurements only
0%
100%
Latency tail detection
20%
50%
Without context: $0.6749 · 3m 26s · 22 turns · 29 in / 13,456 out tokens
With context: $0.5877 · 2m 4s · 21 turns · 478 in / 7,304 out tokens
Adaptive optimizer bottleneck thresholds and protocol-specific tuning
Sort by confidence x improvement
0%
30%
30-second observation wait
0%
100%
Revert threshold
0%
100%
Confidence filter
30%
100%
Raft: max_batch_size recommendation
40%
100%
Byzantine: request_pipelining recommendation
30%
100%
CPU bottleneck threshold
0%
0%
Memory bottleneck threshold
0%
0%
Network bottleneck threshold
0%
0%
Throughput degradation detection
0%
100%
Without context: $0.5765 · 2m 54s · 16 turns · 18 in / 12,893 out tokens
With context: $0.6466 · 2m 19s · 21 turns · 279 in / 9,071 out tokens
b2618f9
Table of Contents
If you maintain this skill, you can claim it as your own. Once claimed, you can manage eval scenarios, bundle related skills, attach documentation or rules, and ensure cross-agent compatibility.