agentmods.dev

cosim-gpu: Skill for Claude Code

.agents/skills/cosim-gpu-rocm-stack/SKILL.md

cosim-gpu-rocm-stack is a skill for Claude Code, Codex from gevico/cosim-gpu. It costs 50 tokens per session (446 once invoked), scanned A, original, Apache-2.0.

A guest-side guide for debugging ROCm behavior in a QEMU and gem5 GPU co-simulation. ROCm is AMD's software stack for GPU computing, including HIP, the HSA runtime, and kernel drivers.

In plain words
What is it for?
Use it to investigate HSA signals, HIP waits, KFD driver calls, device visibility, driver loading, PASID or VMID routing, and related cosimulation failures.
Why use it?
It helps identify whether a problem comes from test settings, the simulated guest, debug evidence, or the GPU model. It also points to the right logs, device checks, and runtime reference material for low-level symptoms.

Skill for Claude CodeCodex

Written for no agent in particular: nothing here depends on one. Also seen: installed under .agents/ (shared by several agents).

This is gevico/cosim-gpu's own configuration. It tells Claude Code and Codex how to work on cosim-gpu itself, so it is not a mod to install elsewhere. Copy it as a starting point and replace the rules that are about this project. Everything cosim-gpu configures →

not rated 95repo +16 changed 11d ago A scan 50 tokens original Apache-2.0
View source ↗ gevico/cosim-gpu gevico.github.io
Reuse

Borrowing it

Nothing to install: this file belongs to gevico/cosim-gpu. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.

Copy the file
curl -O https://raw.githubusercontent.com/gevico/cosim-gpu/main/.agents/skills/cosim-gpu-rocm-stack/SKILL.md
Clone the repo
git clone --depth 1 https://github.com/gevico/cosim-gpu

Made for: Claude Code, Codex.

Wrote 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.

agentmods badge for cosim-gpu-rocm-stack

README.md
[![agentmods](https://agentmods.dev/badge/skills/gevico/cosim-gpu/cosim-gpu-rocm-stack/github.svg)](https://agentmods.dev/skills/gevico/cosim-gpu/cosim-gpu-rocm-stack)
Your own site
<a href="https://agentmods.dev/skills/gevico/cosim-gpu/cosim-gpu-rocm-stack"><img src="https://agentmods.dev/badge/skills/gevico/cosim-gpu/cosim-gpu-rocm-stack/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.

agentmods 80×15 button for cosim-gpu-rocm-stack

Your own site · 80×15
<a href="https://agentmods.dev/skills/gevico/cosim-gpu/cosim-gpu-rocm-stack"><img src="https://agentmods.dev/badge/skills/gevico/cosim-gpu/cosim-gpu-rocm-stack.svg" alt="Reviewed on agentmods" width="80" height="20"></a>
Per session 50 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 446 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. A grade says what 26 rules found in the file — not that it is safe.
Origin original No closer match found in the catalogue.
Token cost

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.

ModelPer sessionOnce invoked
Fable 5.1 $0.00050 $0.00446
Opus 5.5 $0.00020 $0.00178
Sonnet 5.5 $0.00010 $0.00089
Haiku 4.5 $0.00005 $0.00045

Measured 11d ago against content hash 54a10aa03fbd, method: parsed. Prices are Anthropic first-party input rates as of 2026-10-07, from the pricing page.

Security

Grade A, and why

cosim-gpu-rocm-stack 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 11d 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.

.agents/skills/cosim-gpu-rocm-stack/SKILL.md · 48 lines

What it actually says

Cosim ROCm Stack

Use this as the vHost-side ROCm routing guide. It explains when a symptom belongs to test policy, vHost inspection, debug evidence, or model-side review. Detailed runtime semantics live in references/rocm-runtime-reference.md; read that file when debugging HSA signals, HIP API waits, KFD ioctls, PASID, VMID, or driver parameters.

Stack Overview

HIP API
  -> ROCclr
    -> ROCr / HSA runtime
      -> amdgpu KFD
        -> amdgpu DRM
          -> gem5 GPU model through QEMU

Routing

Question First route
Environment defaults or HSA_ENABLE_INTERRUPT cosim-gpu-test, then scripts/host/run_cosim_tests.sh
vHost device visibility, driver load, debugfs cosim-gpu-guest
Non-PASS row, hang, crash, or timeout cosim-gpu-debug
Disk image package or service changes cosim-gpu-disk-image-edit
HSA signal, HIP wait, KFD ioctl, PASID or VMID meaning references/rocm-runtime-reference.md

Evidence To Collect

Use cosim-gpu-guest and cosim-gpu-debug to collect:

  • amdgpu and KFD kernel log excerpts
  • PCI device state and loaded kernel modules
  • ROCm agent and memory visibility
  • KFD debugfs state when signal or queue progress is suspected
  • program identity, interrupt mode, verdict, matrix row, and artifact path

When runtime symbols or debug builds are needed, use repository scripts and cosim-gpu-disk-image-edit; do not embed package-install procedures in this skill.

Files

What ships with it

1 file beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.

Changes

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.

  1. 11d ago Changed · +1 tokens per session 54a10aa03fbd
  2. 19d ago First seen · 48 lines · 49 tokens per session scan A a4572e740d51

Subscribe to this mod's changes

cosim-gpu-rocm-stack is a skill published in the GitHub repository gevico/cosim-gpu (95 stars, last pushed 4d ago), licensed Apache-2.0. It adds 50 tokens to every session and 446 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 2026-10-10.

DebuggingHardware and IoT
Related

Other skills, from other repositories

gke-ai-troubleshooting-tpu-dynamic-slices-monitoring

Monitors, troubleshoots, and manages GKE TPU Dynamic Slices custom resources. Use when checking TPU slice lifecycle states, troubleshooting slice provisioning failures, validating single-slice or multi-slice (JobSet) workload manifests, or safely patching stuck finalizers and disabling the slice controller. Don't use…

google/skills · 107 tokens

gke-ai-troubleshooting-tpu-vbar-oom

Diagnoses and prevents vbarcontrolagent segfaults, out-of-memory (OOM) errors, and TPU device initialization failures on TPU v6e nodes in GKE caused by race conditions during TPU device resets or high-frequency metrics polling. Use when troubleshooting vbarcontrolagent crashes, memory cgroup OOMs in serial console…

google/skills · 125 tokens

doca-flow

DOCA Flow packet steering on NVIDIA NICs/DPUs, linked into a C application: match/action pipes, ports, representors, forwarding, hairpin, counters. Trigger on packet steering, classifier, rule-matching, representor, hairpin or 5-tuple-to-queue questions even when "DOCA Flow" is not named. Also covers pipe validation…

NVIDIA/skills · 224 tokens

competition-firmware-layout

Internal downstream skill for ctf-sandbox-orchestrator. CTF-sandbox workflow for firmware images, partition tables, boot chains, update packages, extracted filesystems, embedded configs, and device-facing trust boundaries. Use when the user asks to unpack firmware, map partition layout, inspect bootloader or init…

zhaoxuya520/reverse-skill · 110 tokens

diagnose-driver-install

Diagnose NVIDIA driver installation failures on DeepOps-managed nodes — nvidia-smi errors, "No devices were found", DKMS build failures, or GPU pods crash-looping. Use before reinstalling anything.

NVIDIA/deepops · 47 tokens

catc-troubleshoot

Catalyst Center troubleshooting workflows - device unreachable investigation, client connectivity issues, interface down analysis, site-wide outage triage, wireless roaming problems, integration with pyATS for CLI-level diagnostics. Use when the device is managed by Catalyst Center and you want its…

automateyournetwork/netclaw · 114 tokens