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.
curl -O https://raw.githubusercontent.com/gevico/cosim-gpu/main/.agents/skills/cosim-gpu-debug/SKILL.mdgit clone --depth 1 https://github.com/gevico/cosim-gpuWrote 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/gevico/cosim-gpu/cosim-gpu-debug)<a href="https://agentmods.dev/skills/gevico/cosim-gpu/cosim-gpu-debug"><img src="https://agentmods.dev/badge/skills/gevico/cosim-gpu/cosim-gpu-debug/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/gevico/cosim-gpu/cosim-gpu-debug"><img src="https://agentmods.dev/badge/skills/gevico/cosim-gpu/cosim-gpu-debug.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 | $0.00077 | $0.02128 |
| Opus 5.5 | $0.00031 | $0.00851 |
| Sonnet 5.5 | $0.00015 | $0.00426 |
| Haiku 4.5 | $0.00008 | $0.00213 |
Grade A, and why
cosim-gpu-debug 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 4d 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 — 192 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Cosim Debug
Use this skill as the debugging entry point for non-PASS cosim behavior. Keep the workflow evidence-first: identify the authoritative artifact, decide the first failing component, then read only the reference that matches the observed mechanism.
Entry Guard
For debug requests, use this workflow directly. Do not search for retired debug skills or older routing names. Translate historical dialogue into concrete evidence fields: artifact path, failing command, program identity, run id, environment row, first durable failure, live wait state, comparison row, and source mechanism under inspection.
Environment variable questions are not debug evidence by themselves. Use
cosim-gpu-test for runner prefix handling and cosim-gpu-rocm-stack for ROCm meaning.
Return here when the environment row produces a non-PASS result or a live wait
state.
Evidence Authority
Use cosim-gpu-info-gathering for broad artifact intake, prior conversation
review, or any task that may otherwise require scanning many logs. This debug
skill should reason from the compact evidence map before opening raw logs.
Start by classifying the artifact state:
| Artifact state | Action |
|---|---|
verdict.json, matrix.tsv, and patch/binary-provenance.txt exist |
Treat archived files as authority. |
| Runner row is incomplete and a matching process is alive | Take bounded live samples only when the active plan allows it. |
| Runner row is incomplete and dead | Rerun only the accepted manifest row through cosim-gpu-test. |
| Old row lacks gem5 log needed for ownership | Create a fixed rerun row with the same program, environment, timeout, and binary. |
| QEMU-visible failure has no matching gem5 evidence | Classify as provisional until gem5 log is available. |
Record the exact source and binary provenance before accepting any build or
test evidence. Use cosim-gpu-build for build rules; do not duplicate its checks
here.
Observable Event Routing
For timeout, wait, and non-PASS rows, prefer a compact event summary before
reading broad logs. Debug flags such as GPUWgProgress,
HSAPacketProcessor, and GPUCommandProc are event sources, not the
classification model. Different logs may fill the same observation dimension.
What ships with it
16 files 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.
- references/analysis/debug-analysis.md 2.5 KB
- references/analysis/debug-workflows.md 3.6 KB
- references/analysis/live-wait-state.md 2.9 KB
- references/analysis/observable-dimensions.md 4.5 KB
- references/gem5-model/address-translation-fault.md 3.0 KB
- references/gem5-model/cache-coherence-checkpoints.md 2.5 KB
- references/gem5-model/discovery-log.md 5.9 KB
- references/gem5-model/examples/signal-completion-debug.md 1.8 KB
- references/gem5-model/examples/vmid-assert-crash.md 1.5 KB
- references/gem5-model/hsa-signal-completion-pattern.md 4.2 KB
- references/gem5-model/overview.md 4.0 KB
- references/gem5-model/vmid-assert-lesson.md 4.4 KB
- references/gem5-model/vmid-pasid-architecture.md 5.4 KB
- references/qemu/error-setv-pattern.md 3.9 KB
- references/qemu/examples/proxy-double-error.md 1.6 KB
- references/qemu/qemu-first-failure.md 1.4 KB
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.
- 4d ago Changed a9b01ccc9e7c
- 11d ago Changed · +1 tokens per session 8a4086b2557b
- 12d ago Changed a3f5601c4e9c
- 19d ago First seen · 192 lines · 76 tokens per session scan A 4ef6f10dc7c0
cosim-gpu-debug is a skill published in the GitHub repository gevico/cosim-gpu (95 stars, last pushed 4d ago), licensed Apache-2.0. It adds 77 tokens to every session and 2,128 once invoked, about $0.0003 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.
Other skills, from other repositories
systematic-debugging
Use when debugging a failing test, build error, or runtime issue that isn't immediately obvious. Guides a 4-phase root cause analysis instead of random fix attempts.
log-error-digest
Analyze log files to troubleshoot errors, identify peak error periods, and produce error clustering, frequency statistics, and time distribution reports. Supports JSON, syslog, and Nginx formats with automatic detection. Use when a user uploads a .log file and asks to analyze errors, find patterns, debug issues, or…
cwv-optimizer
Diagnose and fix Core Web Vitals issues on AEM Edge Delivery Services pages. Goes deeper than generic CWV advice by understanding EDS-specific performance patterns including the 100KB LCP budget, E-L-D loading phases, block rendering behavior, and third-party script impact. Produces specific fixes for LCP, CLS, and…
incident-response-incident-response
Coordinate production incident triage, mitigation, investigation, recovery, and follow-up. Use when responding to outages or other incidents that require structured operational coordination.
ci-fixer
CI failures - read error, minimal fix, verify.
diagnosing-missing-recordings
Diagnoses why a session recording is missing or was not captured. Use when a user asks why a session has no replay, why recordings aren't appearing, or wants to troubleshoot session replay capture issues for a specific session ID or across their project. Covers SDK diagnostic signals, project settings, sampling…