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.
npx skills add yamaru-eu/hardware-probe --skill expert-llmfitgit clone --depth 1 https://github.com/yamaru-eu/hardware-probeWrote 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/skills/yamaru-eu/hardware-probe/expert-llmfit)<a href="https://agentmods.dev/skills/yamaru-eu/hardware-probe/expert-llmfit"><img src="https://agentmods.dev/badge/skills/yamaru-eu/hardware-probe/expert-llmfit/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/skills/yamaru-eu/hardware-probe/expert-llmfit"><img src="https://agentmods.dev/badge/skills/yamaru-eu/hardware-probe/expert-llmfit.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 | $0.00046 | $0.00891 |
| Opus 5 | $0.00023 | $0.00445 |
| Sonnet 5 | $0.00009 | $0.00178 |
| Haiku 4.5 | $0.00005 | $0.00089 |
Grade C, and why
expert-llmfit scanned grade C with 1 finding 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 10d 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.
Recursive force deletehighDestructive command
rm -rf with a variable or a broad path is one typo away from removing the wrong tree.
rm -rf /tmp/llmfit_external How it starts
The opening of the file, as written. The whole thing — 69 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Expert LLMFit (BETA)
This skill provides the domain expertise required to maintain the LlmFitEngine inside remote-affiliate-api/src/core/llmfit.engine.ts and keep it synchronized with the upstream logic from https://github.com/AlexsJones/llmfit.
Core Concepts & Formulas
-
Inference is Memory-Bandwidth-Bound:
- Theoretical Max TPS =
Memory Bandwidth (GB/s) / Model Size (GB) - Real TPS Estimate =
Max TPS * Efficiency Factor (0.55)*RunMode Penalty. - The efficiency factor (0.55) accounts for kernel launch overhead and KV-cache reads.
- Theoretical Max TPS =
-
RunMode Penalties:
gpu: 1.0 (Full speed)moe_offload: 0.8 (Expert switching latency)cpu_gpu: 0.5 (Partial offload penalty)cpu_only: 0.3 (Fallback to slow RAM)
-
Memory Requirements (VRAM/RAM):
- Base Model Size =
Parameters * BytesPerParam- FP16/BF16: 2.0 bytes
- Q8: 1.0 byte
- Q4_K_M: 0.5 bytes (Standard for local inference recommendations)
- KV Cache Overhead: Context length requires additional VRAM.
- Rule of Thumb VRAM:
(Model Size in GB * 1.2)to account for context and OS overhead.
- Base Model Size =
-
Mixture of Experts (MoE):
- Fit/VRAM checks use Total Parameters.
- Speed/TPS estimates use Active Parameters (e.g., Mixtral 8x7B uses ~12.9B active params per token).
Verification & Adaptation Workflow
Whenever asked to "verify LLMFit logic" or "update the hardware recommendation engine", follow these steps:
Step 1: Upstream Sync
- Clone the upstream repository into a temporary directory:
mkdir -p /tmp/llmfit_external && git clone https://github.com/AlexsJones/llmfit /tmp/llmfit_external - Inspect key Rust files:
llmfit-core/src/fit.rs(for TPS and scoring formulas).llmfit-core/src/hardware.rs(for GPU bandwidth updates and compute capabilities).llmfit-core/src/models.rs(for Quantization multipliers and MoE active parameters).
Step 2: Codebase Audit (remote-affiliate-api)
- Compare the upstream formulas with our local TypeScript implementation in
src/core/llmfit.engine.ts. - Ensure the
estimateTokensPerSecondmethod correctly applies the0.55efficiency factor andRunModepenalties. - Verify
calculateRequiredMemoryproperly uses active vs total parameters.
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.
- 10d ago First seen · 69 lines · 46 tokens per session scan C d06a3551bb13
expert-llmfit is a skill published in the GitHub repository yamaru-eu/hardware-probe (7 stars, last pushed 1mo ago), licensed Apache-2.0. It adds 46 tokens to every session and 891 once invoked, about $0.0002 per session on Opus 5. A static security scan graded it C with 1 finding (recursive force delete). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other skills, from other repositories
spark-environment-setup
Set up a working ML training/inference environment on NVIDIA DGX Spark (GB10, aarch64, CUDA 13). Use when installing PyTorch/Unsloth/TRL/vLLM on DGX Spark, hitting libcudart or wheel-ABI errors on aarch64, or choosing between NGC containers and bare pip installs.
spark-memory-thermal-ops
Manage unified memory and thermals during long-running ML jobs on NVIDIA DGX Spark. Use when planning memory headroom for a training run on GB10, when a job OOMs on unified memory, or when monitoring temperature and power during multi-hour training.
spark-training-gotchas
Preflight and diagnose the ten known failure modes for ML training on NVIDIA DGX Spark. Use when a training run on DGX Spark fails to start, OOMs below the 128GB limit, slows down mid-run, or before any multi-hour training job on GB10.
llama-cpp
Runs LLM inference on CPU, Apple Silicon, and consumer GPUs without NVIDIA hardware. Use for edge deployment, M1/M2/M3 Macs, AMD/Intel GPUs, or when CUDA is unavailable. Supports GGUF quantization (1.5-8 bit) for reduced memory and 4-10× speedup vs PyTorch on CPU.
amc-run-rtsp-calibration
Calibrate a new dataset from live RTSP camera streams via the AutoMagicCalib REST API. Use when the user provides RTSP URLs or asks to calibrate live cameras; VIOS records clips, AMC ingests them, then runs calibration.
deepstream-import-vision-model
Use this skill to bring a supported object-detection vision model from HuggingFace or NVIDIA NGC into an NVIDIA DeepStream pipeline with end-to-end automation: ONNX download, SafeTensors export, TRT engine build, custom nvinfer bbox parser, multi-stream benchmark, and PDF report. Object detection models only.