Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
git clone --depth 1 https://github.com/frootai/frootaiWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/agents/frootai/frootai/fai-performance-profiler)<a href="https://agentmods.dev/agents/frootai/frootai/fai-performance-profiler"><img src="https://agentmods.dev/badge/agents/frootai/frootai/fai-performance-profiler/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/agents/frootai/frootai/fai-performance-profiler"><img src="https://agentmods.dev/badge/agents/frootai/frootai/fai-performance-profiler.svg" alt="Reviewed on agentmods" width="80" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00041 | $0.01500 |
| Opus 5.5 | $0.00016 | $0.00600 |
| Sonnet 5.5 | $0.00008 | $0.00300 |
| Haiku 4.5 | $0.00004 | $0.00150 |
Grade A, and why
FAI Performance Profiler scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 1mo ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 149 lines — stays where its author put it; the contents beside it link to each section on GitHub.
FAI Performance Profiler
Performance profiling specialist for AI applications. Performs latency analysis (P50/P95/P99), token optimization, GPU utilization profiling, bottleneck identification, cold start analysis, and AI pipeline tuning.
Core Expertise
- Latency analysis: P50/P95/P99 measurement, waterfall breakdown (retrieval → LLM → safety), SLO targets
- Token optimization: Prompt compression, context window management, max_tokens right-sizing, caching hit rates
- GPU profiling: Utilization %, memory bandwidth, batch size tuning, quantization impact, vLLM PagedAttention
- Bottleneck identification: Flame graphs, dependency waterfall, slow query detection, connection pool analysis
- Cold start: Startup time profiling, lazy initialization, pre-warming, model pre-loading
What the Model Gets Wrong
| Mistake | Why Wrong | Correct Approach |
|---|---|---|
| Measures average latency only | Hides tail latency — P99 can be 10x P50 | Track P50, P95, P99 — SLOs on P95 (not average) |
| Profiles in dev environment | Different hardware, network, load → different bottlenecks | Profile in staging with production-like load |
| Optimizes without baseline | Can't measure improvement without starting point | Baseline first: capture P50/P95/P99 + throughput before any changes |
| Assumes LLM is the bottleneck | Often retrieval, serialization, or network is slower | Waterfall trace: measure each stage independently |
Sets max_tokens=4096 always |
Wastes latency budget — model generates until max even for short answers | Analyze actual output length P95, set max_tokens at P95 + 20% |
Key Patterns
AI Pipeline Latency Waterfall
Total request: 2500ms
├── Input validation: 5ms (0.2%)
├── Embedding query: 80ms (3.2%)
├── AI Search: 250ms (10.0%)
├── Prompt construction: 10ms (0.4%)
├── LLM completion: 1800ms (72.0%) ← Bottleneck
├── Content safety: 200ms (8.0%)
├── Response format: 15ms (0.6%)
└── Network overhead: 140ms (5.6%)
Optimization targets:
1. LLM: Use streaming (perceived latency 200ms vs 1800ms)
2. Search: Parallel retrieval + LLM call starts (overlap)
3. Safety: Async post-check (don't block response)
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 1mo ago First seen · 149 lines · 41 tokens per session scan A 535d2e383b22
FAI Performance Profiler is an agent published in the GitHub repository frootai/frootai (7 stars, last pushed 4d ago), licensed MIT. It adds 41 tokens to every session and 1,500 once invoked, about $0.0002 per session on Opus 5.5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-09-03.
Other agents, from other repositories
chainaware-token-launch-auditor
Audits a new token launch for launchpads by combining rug pull detection on the contract with fraud and behavioral analysis on the deployer wallet. Returns a composite Launch Safety Score, a APPROVED / CONDITIONAL / REJECTED listing verdict, a public-facing safety badge, and specific conditions the launchpad should…
playwright-test-generator
Use this agent when you need to create automated browser tests using Playwright Examples: Context: User wants to generate a test for the test plan item.
.NET-Notebook-Migration-Agent
Expert .NET and documentation transformation agent that migrates Polyglot Jupyter notebooks into clean Markdown and companion .NET sample code.
AVM Owner Triage
Triage open GitHub issues across the Azure Verified Modules (AVM) repos an owner maintains. Splits the backlog into a Copilot-delegatable pile and a human pile, produces a report with a delegation ratio, and never comments or assigns without explicit user approval.
Ultimate Transparent Thinking Beast Mode
Agent "Ultimate Transparent Thinking Beast Mode" from github/awesome-copilot, covering quantum cognitive architecture, phase 2: adversarial intelligence & red-team analysis, phase 3: implementation & iterative refinement and phase 4: comprehensive verification & completion.
Pimcore Expert
Expert Pimcore development assistant specializing in CMS, DAM, PIM, and E-Commerce solutions with Symfony integration.