Agent skill
by jeremylongshore in jeremylongshore/tons-of-skills-marketplace
Triage a dead or degraded GPU on a CoreWeave node fast — decide reschedule vs GPU-reset vs node-reboot vs RMA from an Xid code or a pasted dmesg / nvidia-smi blob, so a bad card does not silently…
$ npx skills add jeremylongshore/tons-of-skills-marketplace --skill coreweave-gpu-node-forensics -a claude-codeProject install by default; add -g for ~/.claude/skills/.
$ gh skill install jeremylongshore/tons-of-skills-marketplace coreweave-gpu-node-forensics --agent claude-codeProject scope by default; add --scope user for a personal install. Needs GitHub CLI 2.90.0 or later (public preview).
$ git clone --depth 1 https://github.com/jeremylongshore/tons-of-skills-marketplace.git skills-src && mkdir -p .claude/skills && cp -r skills-src/skills/.curated/coreweave-gpu-node-forensics .claude/skills/coreweave-gpu-node-forensics && rm -rf skills-srcUse ~/.claude/skills/ instead of .claude/skills for a personal install. The folder must contain SKILL.md.
Claude Code skills documentation · loads skills from .claude/skills/
Install the "coreweave-gpu-node-forensics" agent skill from https://github.com/jeremylongshore/tons-of-skills-marketplace/tree/main/skills/.curated/coreweave-gpu-node-forensics into .claude/skills/coreweave-gpu-node-forensics/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "coreweave-gpu-node-forensics", then confirm the skill loads.Claude Code copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$skill-installer install https://github.com/jeremylongshore/tons-of-skills-marketplace/tree/main/skills/.curated/coreweave-gpu-node-forensicsType this inside Codex. $skill-installer <name> installs a curated skill from openai/skills. The installer writes to $CODEX_HOME/skills (default ~/.codex/skills). Restart Codex if the skill does not show up.
$ npx skills add jeremylongshore/tons-of-skills-marketplace --skill coreweave-gpu-node-forensics -a codexProject install goes to .agents/skills/; add -g for ~/.codex/skills/.
$ gh skill install jeremylongshore/tons-of-skills-marketplace coreweave-gpu-node-forensics --agent codexProject scope by default (.agents/skills/); add --scope user for a personal install.
$ git clone --depth 1 https://github.com/jeremylongshore/tons-of-skills-marketplace.git skills-src && mkdir -p .agents/skills && cp -r skills-src/skills/.curated/coreweave-gpu-node-forensics .agents/skills/coreweave-gpu-node-forensics && rm -rf skills-srcUse ~/.agents/skills/ instead of .agents/skills for a personal install.
Codex skills documentation · loads skills from .agents/skills/
Install the "coreweave-gpu-node-forensics" agent skill from https://github.com/jeremylongshore/tons-of-skills-marketplace/tree/main/skills/.curated/coreweave-gpu-node-forensics into .agents/skills/coreweave-gpu-node-forensics/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "coreweave-gpu-node-forensics", then confirm the skill loads.Codex copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ npx skills add jeremylongshore/tons-of-skills-marketplace --skill coreweave-gpu-node-forensics -a cursorProject install goes to .agents/skills/; add -g for ~/.cursor/skills/.
$ gh skill install jeremylongshore/tons-of-skills-marketplace coreweave-gpu-node-forensics --agent cursorProject scope by default (.agents/skills/); add --scope user for a personal install.
$ git clone --depth 1 https://github.com/jeremylongshore/tons-of-skills-marketplace.git skills-src && mkdir -p .cursor/skills && cp -r skills-src/skills/.curated/coreweave-gpu-node-forensics .cursor/skills/coreweave-gpu-node-forensics && rm -rf skills-srcUse ~/.cursor/skills/ instead of .cursor/skills for a personal install.
Cursor skills documentation · loads skills from .cursor/skills/, .agents/skills/, .claude/skills/, .codex/skills/
Install the "coreweave-gpu-node-forensics" agent skill from https://github.com/jeremylongshore/tons-of-skills-marketplace/tree/main/skills/.curated/coreweave-gpu-node-forensics into .cursor/skills/coreweave-gpu-node-forensics/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "coreweave-gpu-node-forensics", then confirm the skill loads.Cursor copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ gemini skills install https://github.com/jeremylongshore/tons-of-skills-marketplace.git --path skills/.curated/coreweave-gpu-node-forensics--scope user (default) or --scope workspace; --path is the subfolder of the repo that holds the skill; --consent skips the security confirmation prompt.
$ npx skills add jeremylongshore/tons-of-skills-marketplace --skill coreweave-gpu-node-forensics -a gemini-cliProject install goes to .agents/skills/; add -g for ~/.gemini/skills/.
$ gh skill install jeremylongshore/tons-of-skills-marketplace coreweave-gpu-node-forensics --agent gemini-cliProject scope by default (.agents/skills/); add --scope user for a personal install.
$ git clone --depth 1 https://github.com/jeremylongshore/tons-of-skills-marketplace.git skills-src && mkdir -p .gemini/skills && cp -r skills-src/skills/.curated/coreweave-gpu-node-forensics .gemini/skills/coreweave-gpu-node-forensics && rm -rf skills-srcUse ~/.gemini/skills/ instead of .gemini/skills for a personal install, then run /skills reload.
Gemini CLI skills documentation · loads skills from .gemini/skills/, .agents/skills/
Install the "coreweave-gpu-node-forensics" agent skill from https://github.com/jeremylongshore/tons-of-skills-marketplace/tree/main/skills/.curated/coreweave-gpu-node-forensics into .gemini/skills/coreweave-gpu-node-forensics/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "coreweave-gpu-node-forensics", then confirm the skill loads.Gemini CLI copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ gh skill install jeremylongshore/tons-of-skills-marketplace coreweave-gpu-node-forensicsInstalls for Copilot at project scope by default; add --scope user for a personal install. Preview a skill first with gh skill preview. Needs GitHub CLI 2.90.0 or later (public preview).
$ npx skills add jeremylongshore/tons-of-skills-marketplace --skill coreweave-gpu-node-forensics -a github-copilotProject install goes to .agents/skills/; add -g for ~/.copilot/skills/.
$ git clone --depth 1 https://github.com/jeremylongshore/tons-of-skills-marketplace.git skills-src && mkdir -p .github/skills && cp -r skills-src/skills/.curated/coreweave-gpu-node-forensics .github/skills/coreweave-gpu-node-forensics && rm -rf skills-srcUse ~/.copilot/skills/ instead of .github/skills for a personal install. Commit .github/skills so cloud agent and code review can use it.
GitHub Copilot skills documentation · loads skills from .github/skills/, .claude/skills/, .agents/skills/
Install the "coreweave-gpu-node-forensics" agent skill from https://github.com/jeremylongshore/tons-of-skills-marketplace/tree/main/skills/.curated/coreweave-gpu-node-forensics into .github/skills/coreweave-gpu-node-forensics/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "coreweave-gpu-node-forensics", then confirm the skill loads.GitHub Copilot copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ npx skills add jeremylongshore/tons-of-skills-marketplace --skill coreweave-gpu-node-forensics -a opencodeOpenCode documents no install command of its own. Project install goes to .agents/skills/; add -g for ~/.config/opencode/skills/.
$ gh skill install jeremylongshore/tons-of-skills-marketplace coreweave-gpu-node-forensics --agent opencodeProject scope by default (.agents/skills/); add --scope user for a personal install.
$ git clone --depth 1 https://github.com/jeremylongshore/tons-of-skills-marketplace.git skills-src && mkdir -p .opencode/skills && cp -r skills-src/skills/.curated/coreweave-gpu-node-forensics .opencode/skills/coreweave-gpu-node-forensics && rm -rf skills-srcUse ~/.config/opencode/skills/ instead of .opencode/skills for a personal install.
OpenCode skills documentation · loads skills from .opencode/skills/, .claude/skills/, .agents/skills/
Install the "coreweave-gpu-node-forensics" agent skill from https://github.com/jeremylongshore/tons-of-skills-marketplace/tree/main/skills/.curated/coreweave-gpu-node-forensics into .opencode/skills/coreweave-gpu-node-forensics/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "coreweave-gpu-node-forensics", then confirm the skill loads.OpenCode copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
coreweave-gpu-node-forensicsTriage a dead or degraded GPU on a CoreWeave node fast — decide reschedule vs GPU-reset vs node-reboot vs RMA from an Xid code or a pasted dmesg / nvidia-smi blob, so a bad card does not silently…
Coreweave GPU Node Forensics is an agent skill from jeremylongshore/tons-of-skills-marketplace. Triage a dead or degraded GPU on a CoreWeave node fast — decide reschedule vs GPU-reset vs node-reboot vs RMA from an Xid code or a pasted dmesg / nvidia-smi blob, so a bad card does not silently kill a multi-day training run. Use when a GPU throws an Xid error, a node "fell off the bus", a training run stalls or NCCL hangs on one rank, or you need to know whether to replace, reset, or just reschedule. Trigger with "xid error", "gpu fell off the bus", "coreweave gpu dead", "should I RMA this GPU", "gpu node…
Its SKILL.md is about 3.2k tokens, which your agent loads only when the skill is triggered. The skill folder holds 9 other files, including scripts and reference files (for example `ARD.md`, `PRD.md` and `eval-spec.yaml`). Compatibility notes: Designed for Claude Code
It sits in AI & LLM Engineering, covering GPU and accelerator computing. It works with NVIDIA AI Platform. The repository describes itself as: Model-agnostic agent-skills platform with a harness-free canonical layer, verified adapters, and the ccpi package manager. Explore at tonsofskills.com. The licence is MIT.
6 steps, taken from the step headings in SKILL.md.
Read from SKILL.md and the folder at commit 23ea8d4. It shows what the files ask for, not the result of running them.
Pre-approves these tools, so the agent can use them without asking each time:
ReadBash(python3:*)Bash(nvidia-smi -q:*)Bash(kubectl get:*)Bash(dmesg:*)From allowed-tools in the SKILL.md frontmatter.
Ships 1 file in scripts/ (Python), which the agent can run.
Shell commands in SKILL.md call:
python3kubectljqFrom the folder's file list and the shell code blocks in SKILL.md.
Links to these hosts (documentation or services it may open):
docs.nvidia.comFrom URLs in SKILL.md, links to its own repository left out.
Names no API keys, tokens, secrets or passwords.
From names ending in _API_KEY, _TOKEN, _SECRET, _KEY or _PASSWORD in SKILL.md.
Designed for Claude Code
From compatibility in the SKILL.md frontmatter.
Coreweave GPU Node Forensics loads about 3.2k tokens when it runs, and up to ~6.2k if it reads all its reference files. Until then it costs about 138 tokens; SKILL.md has 1,418 words of instructions outside code blocks.
Estimates: characters ÷ 4, the usual rule of thumb; real counts depend on the model's tokenizer. Scripts and assets cost tokens only if the agent reads them.
The automated check found no risky patterns in SKILL.md.
Automated static check — not a guarantee. Review scripts before installing. It scans the text of SKILL.md for risky patterns (piping downloads into a shell, reading credential files, hidden Unicode, destructive commands); the scripts in this folder are not scanned.
The full file from jeremylongshore/tons-of-skills-marketplace at commit 23ea8d4, republished under its MIT licence (© jeremylongshore). 1,418 words, ~3,207 tokens.
.claude/skills/coreweave-gpu-node-forensics/SKILL.md (or your agent's skills folder). This skill also uses 7 other files; get the full folder from GitHub.Community-contributed. Not affiliated with, endorsed by, or sponsored by CoreWeave, Inc. CoreWeave is a registered trademark of CoreWeave, Inc. "NVIDIA" and "Xid" are trademarks of NVIDIA Corporation; Xid semantics are cited from NVIDIA's public documentation.
Triages a dead or degraded GPU on a CoreWeave node in seconds and returns one
grounded move — reschedule, reset-gpu, reboot-node, rma, watch, or
app-bug-not-hardware — from an NVIDIA Xid code or a pasted dmesg /
nvidia-smi blob. The classification is deterministic: a bundled script does the
mapping so the agent never guesses whether a card is dead, degraded, or fine.
A single bad GPU can kill a 64-GPU, multi-day training run — thousands of dollars and days of wall-clock gone — because one rank stalls the whole collective. The expensive mistakes are triage mistakes: RMAing a healthy card for an app bug, restarting a job onto a GPU whose memory error was uncontained, or manually uncordoning a node the lifecycle controller is trying to replace. This skill kills that ambiguity.
The decision logic is grounded in the NVIDIA Xid error catalog
(https://docs.nvidia.com/deploy/xid-errors/) and CoreWeave's node-lifecycle /
cordon behavior. The math-of-the-matter — which Xid means what, and how
row-remapper state overrides it — lives in scripts/triage.py as a table the LLM
does not get to re-litigate. Deep domain knowledge (the full code→action table,
the row-remap decision, the cordon rules) loads from references/ on demand.
The headline is the Xid 94-vs-95 split. A contained memory error (94) cost you one job restart on a healthy node; an uncontained one (95) means the GPU could not isolate the fault and everything it touched is suspect. Getting that one bit wrong is the difference between a 30-second reschedule and a run that quietly trained on corrupt gradients. The script decides it; the skill never eyeballs it.
This skill is diagnostic, not destructive: it recommends the cordon / drain /
reset / RMA next-step but its tools are scoped read-only (nvidia-smi -q,
kubectl get, dmesg) — it never runs a reset, a reboot, or an uncordon itself.
dmesg /
nvidia-smi dump. The skill works from a paste alone — no live cluster access
required — which is the common case (an operator pastes what the run's logs
showed).kubectl context on the CoreWeave
cluster (read-only is enough), and nvidia-smi on the node. If neither is
available the skill still triages from the paste.python3 to run the deterministic classifier (scripts/triage.py, stdlib
only — no dependencies).No secrets are handled. All commands are read-only queries.
The pipeline is capture → classify → act. The classifier is authoritative for
the verdict; references/ supplies the "why" when a case needs depth.
If the user has not already pasted it, ask for (or read) the fault signal. The two richest sources:
dmesg -T | grep -i xid # the Xid line(s) with timestamps
nvidia-smi -q -d ROW_REMAPPER,ECC,PERFORMANCE # remap state, ECC counts, throttle reasonsOn a CoreWeave node you can also check who owns any cordon before acting:
kubectl get node NODE -o json | jq '{unschedulable: .spec.unschedulable, taints: .spec.taints}'Feed the Xid code, or the whole blob, to the classifier. Do not classify by hand — the script owns the Xid→action mapping and the row-remap override.
# From an Xid code:
python3 "${CLAUDE_SKILL_DIR}/scripts/triage.py" --xid 95
# With row-remapper state (Xid 63/64 or a DBE):
python3 "${CLAUDE_SKILL_DIR}/scripts/triage.py" --xid 63 --pending yes
python3 "${CLAUDE_SKILL_DIR}/scripts/triage.py" --xid 48 --remap-failure yes
# From a pasted dmesg / nvidia-smi blob (file or stdin):
python3 "${CLAUDE_SKILL_DIR}/scripts/triage.py" --blob /path/to/dmesg.txt
dmesg -T | grep -i xid | python3 "${CLAUDE_SKILL_DIR}/scripts/triage.py"The script emits a VERDICT block plus a JSON object
({classification, severity, action, why, next_command, cordon_rule?, unverified?}).
Add --json for machine-readable output only.
When a blob carries several Xids, the most severe one governs the move and the
rest are reported as co_occurring_xids — a co-occurring app-side Xid 43 next to a
hardware Xid 79 does not soften the "reboot the node" verdict.
Present the action and the next_command to the user in plain language. The six
actions and what each means live in
${CLAUDE_SKILL_DIR}/references/xid-triage-table.md.
Load it when the user wants the full table or asks about an Xid not in the summary.
reschedule — restart the failed rank; the node stays in service.reset-gpu — drain the GPU and reset it; re-run any job that shared it.reboot-node — the card is off the bus; only a bare-metal reboot returns it.rma — terminal hardware fault; the card must be replaced.watch — correctable / trending; monitor, do not act yet.app-bug-not-hardware — the app faulted, not the GPU. Do NOT RMA.If the Xid is 94 or 95, state the containment explicitly, because the two look almost identical in the logs and lead to opposite actions:
reschedule. The error was isolated to the app's
context; the GPU and node are healthy. Restart the job. Do not reset or RMA.reset-gpu. The GPU could not isolate it; every
context it touched is suspect. Drain, reset, and re-run anything that shared the
card — otherwise the run may continue on corrupt state.For any ECC/DBE or row-remap Xid, the nvidia-smi -q -d ROW_REMAPPER fields
override the base verdict. Pass them in (--pending, --remap-failure) and let
the script decide. The rule:
Remapping Failure Occurred: Yes → rma (terminal — sparing failed).Pending: Yes (no failure) → reset-gpu (routine — applies on reset).Full logic + how to read the histogram:
${CLAUDE_SKILL_DIR}/references/row-remap-decision.md.
Whenever the action is a hardware move (reset-gpu, reboot-node, rma), the
verdict carries a cordon_rule. Surface it verbatim:
Never manually uncordon a CoreWeave health cordon — the node-lifecycle controller owns cordon/uncordon and is driving remediation. Uncordoning re-admits the run onto known-bad hardware and races the controller.
Determine cordon provenance before touching schedulability. Full guidance:
${CLAUDE_SKILL_DIR}/references/cordon-rules.md.
{classification, severity, action, why, next_command} for the governing signal, presented in plain language with the
single next command an operator runs.[unverified] hedges surfaced honestly — e.g. CoreWeave's exact auto-RMA
thresholds and health-cordon taint strings are not publicly published, and the
skill says so rather than inventing a number.| Situation | Cause | Response |
|---|---|---|
| Unmapped Xid | Xid not in the triage table | Script returns a conservative reset-gpu default, flags it [unverified], and points at the NVIDIA catalog. Do not RMA on an unmapped Xid. |
| Xid 94 vs 95 ambiguity | Both are "ECC error" in casual logs | Never infer from prose — use the exact code. 94 = contained (reschedule), 95 = uncontained (reset). |
| Row-remap without an Xid | Only nvidia-smi pasted | Script reads Remapping Failure Occurred / Pending and decides from those alone. |
| App-side Xid mistaken for hardware | Xid 13/31/43/45 | Classed app-bug-not-hardware; the fix is compute-sanitizer on the job, never an RMA. |
| Multiple Xids in one blob | Cascade (e.g. 43 then 79) | Most-severe governs; others listed as co_occurring_xids. |
| Empty / no signal | Blob has no Xid, remap, or thermal line | Script returns watch and asks for fresh dmesg + nvidia-smi -q evidence. |
| Thermal throttle, not a fault | HW/SW Thermal Slowdown : Active | Classed watch — healthy but hot; investigate cooling, not the card. |
| Tempted to uncordon a stalled node | Health cordon owned by the controller | Refuse. Surface the cordon rule; let remediation run. |
VERDICT: Xid 94 — Contained ECC/memory error [severity: HIGH]
ACTION: reschedule
WHY: The error was CONTAINED to the faulting application's context — the GPU
and node are healthy. Just reschedule/restart the job; no reset or RMA needed.
NEXT: Restart or reschedule the failed rank; the node stays in service.No replacement. The containment did its job — restart the rank and keep the node.
VERDICT: Xid 95 — Uncontained ECC/memory error [severity: CRITICAL]
ACTION: reset-gpu
WHY: The error was UNCONTAINED — the GPU could not isolate it, so every context
it touched is suspect. Drain the GPU and reset it; treat all in-flight work as corrupt.
NEXT: Cordon + drain, GPU-reset (nvidia-smi -r) or node reset; re-run any job that shared this GPU.
CORDON: Do NOT manually uncordon a CoreWeave health cordon — the node-lifecycle
controller owns cordon/uncordon. Let it drain and replace the node.Opposite of Example 1 despite looking identical in the logs: drain, reset, and re-run the shared work — do not just restart.
python3 scripts/triage.py --xid 79 → reboot-node. The card is unreachable on
the PCIe bus and returns only after a bare-metal reboot; expect the controller to
cordon and reboot — let it, and RMA only if it recurs after the reboot.
python3 scripts/triage.py --xid 63 --remap-failure yes → rma. The remapper
tried to swap in a spare row and physically could not — terminal. Cordon, drain,
open the RMA. (The script flags that CoreWeave's exact auto-RMA threshold is
[unverified].)
python3 scripts/triage.py --xid 43 → app-bug-not-hardware. Software-induced
fault — the fix is in the application (run it under compute-sanitizer), not an RMA.
Pulling the card would waste healthy hardware and never fix the job.
${CLAUDE_SKILL_DIR}/references/xid-triage-table.md — the full Xid→action table, the six actions, the 94-vs-95 distinction, cited to the NVIDIA Xid catalog.${CLAUDE_SKILL_DIR}/references/row-remap-decision.md — the ROW_REMAPPER routine-vs-terminal decision and how to read the histogram.${CLAUDE_SKILL_DIR}/references/cordon-rules.md — CoreWeave health-cordon ownership and the safe operator path per action.© jeremylongshore, MIT. Rendered from Markdown: HTML in the file is shown as text, images as links, and headings moved down two levels. Raw file
SKILL.md and 7 other files (scripts, references) in skills/.curated/coreweave-gpu-node-forensics of jeremylongshore/tons-of-skills-marketplace.
Open the folder on GitHubat commit 23ea8d4
Coreweave GPU Node Forensics next to the 5 skills that share the most tags, products or categories with it. Stars are the repository's; “used in” counts other GitHub owners with a copy.
| Skill | Stars | Used in | Tokens | Auto-check | Licence | Repo updated |
|---|---|---|---|---|---|---|
| Coreweave GPU Node Forensics this skilljeremylongshore/tons-of-skills-marketplace | 2.8k | — | ~3.2k | Automated safety check: Pass | MIT | |
| Fla Triton To Gluonfla-org/flash-linear-attention | 5.8k | — | ~4.2k | Automated safety check: Pass | MIT | |
| DGX Spark Memory and Thermal Opswshobson/agents | 40k | 1 repos | ~2k | Automated safety check: Pass | MIT | |
| DGX Spark Training Gotchaswshobson/agents | 40k | 1 repos | ~2k | Automated safety check: Pass | MIT | |
| Megatron-LM on SLURMNVIDIA/Megatron-LM | 18k | — | ~1.8k | Automated safety check: Pass | Apache-2.0 | |
| Physical AI Video Augmentation on OSMONVIDIA/skills | 3.5k | — | ~4.7k | Automated safety check: Notes | Apache-2.0 |
fla-org/flash-linear-attention
Workflow for porting an existing Triton kernel in fla/ops/ to Gluon (triton.experimental.gluon) to gain explicit control over tensor layouts, shared memory, async data movement (cp.async / TMA), MMA…
wshobson/agents
Plans memory headroom, works through out-of-memory failures and watches temperature and power during long ML training jobs on NVIDIA DGX Spark.
wshobson/agents
Preflight checks and diagnosis for ten known failure modes of ML training on NVIDIA DGX Spark's GB10, spanning launch errors, memory, thermals, bandwidth and precision.
NVIDIA/Megatron-LM
Shows how to launch distributed Megatron-LM training on a SLURM cluster: sbatch skeleton, torch.distributed.run setup, CUDA_DEVICE_MAX_CONNECTIONS rules and failure diagnosis.
NVIDIA/skills
Orchestrates video data augmentation and auto-labeling workflows on OSMO, from flow selection and preflight checks to submission, monitoring and output download.
Orchestra-Research/AI-Research-SKILLs
Fine-tunes and evaluates OpenVLA-OFT and OFT+ robot policies with LoRA and continuous action heads on LIBERO simulation and ALOHA real-robot setups.
jeremylongshore/tons-of-skills-marketplace
Execute this skill enables AI assistant to conduct a security-focused code review using the security-agent plugin.
jeremylongshore/tons-of-skills-marketplace
Execute this skill enables AI assistant to perform natural language processing and text analysis using the nlp-text-analyzer plugin.
jeremylongshore/tons-of-skills-marketplace
Execute this skill allows AI assistant to construct and configure neural network architectures using the neural-network-builder plugin.
jeremylongshore/tons-of-skills-marketplace
Process identify anomalies and outliers in datasets using machine learning algorithms.
jeremylongshore/tons-of-skills-marketplace
Build this skill enables AI assistant to provide interpretability and explainability for machine learning models.
jeremylongshore/tons-of-skills-marketplace
Execute this skill optimizes prompts for large language models (llms) to reduce token usage, lower costs, and improve performance.
Works with
Categories
Triage a dead or degraded GPU on a CoreWeave node fast — decide reschedule vs GPU-reset vs node-reboot vs RMA from an Xid code or a pasted dmesg / nvidia-smi blob, so a bad card does not silently…. Coreweave GPU Node Forensics is an agent skill from jeremylongshore/tons-of-skills-marketplace. Triage a dead or degraded GPU on a CoreWeave node fast — decide reschedule vs GPU-reset vs node-reboot vs RMA from an Xid code or a pasted dmesg / nvidia-smi blob, so a bad card does not silently kill a multi-day training run.
Coreweave GPU Node Forensics fits situations like: A GPU throws an Xid error; A node fell off the bus; A training run stalls; NCCL hangs on one rank.
Run `npx skills add jeremylongshore/tons-of-skills-marketplace --skill coreweave-gpu-node-forensics -a claude-code`. Or copy the skill folder (skills/.curated/coreweave-gpu-node-forensics in jeremylongshore/tons-of-skills-marketplace) into .claude/skills/coreweave-gpu-node-forensics in your project. Claude Code loads it when a task matches its description.
Run `npx skills add jeremylongshore/tons-of-skills-marketplace --skill coreweave-gpu-node-forensics -a codex`. Or copy the skill folder (skills/.curated/coreweave-gpu-node-forensics in jeremylongshore/tons-of-skills-marketplace) into .agents/skills/coreweave-gpu-node-forensics in your project. Codex loads it when a task matches its description.
Cursor, Gemini CLI, GitHub Copilot and OpenCode also load SKILL.md folders. With the skills CLI, run `npx skills add jeremylongshore/tons-of-skills-marketplace --skill coreweave-gpu-node-forensics -a cursor` (or -a gemini-cli, github-copilot or opencode for the others). To copy it by hand, put the folder in .cursor/skills/coreweave-gpu-node-forensics, .gemini/skills/coreweave-gpu-node-forensics, .github/skills/coreweave-gpu-node-forensics and .opencode/skills/coreweave-gpu-node-forensics in your project.
Going by SKILL.md and its folder, Coreweave GPU Node Forensics needs Python for the scripts in its folder and the command-line tools its instructions call (python3, kubectl and jq). Our summary lists: Python 3. Its frontmatter pre-approves these tools: Read, Bash(python3:*), Bash(nvidia-smi -q:*), Bash(kubectl get:*), Bash(dmesg:*). Compatibility (from SKILL.md): Designed for Claude Code.
SKILL.md names 1 domain. As links in the text: docs.nvidia.com. This is read from the text; nothing was executed.
Our automated static check of SKILL.md found no risky patterns, such as piping downloads into a shell, reading credential files or hidden Unicode. It is not a guarantee. The check reads SKILL.md only: the scripts in the folder are not scanned, so read them before running anything.
Coreweave GPU Node Forensics is published under the MIT licence (declared in SKILL.md). It allows redistribution, so the full SKILL.md is shown on this page.
About 3.2k tokens (SKILL.md is roughly 13k characters). Agents keep only the skill's name and description in context until a task matches; then they load SKILL.md in full. Its references folder adds about 3k tokens, read only when the agent opens those files.
Skills that share tags, products or a category with Coreweave GPU Node Forensics: Fla Triton To Gluon (fla-org/flash-linear-attention, 5.8k stars), DGX Spark Memory and Thermal Ops (wshobson/agents, 40k stars), DGX Spark Training Gotchas (wshobson/agents, 40k stars) and Megatron-LM on SLURM (NVIDIA/Megatron-LM, 18k stars). The comparison table on this page puts their stars, adoption, token cost, safety result and licence side by side.
jeremylongshore (a GitHub user) maintains it in jeremylongshore/tons-of-skills-marketplace, which has 2,821 GitHub stars. The repository holds 3,342 skills in this directory. The repository was last updated on October 8, 2026.
Source: jeremylongshore/tons-of-skills-marketplace on GitHub. Facts on this page come from the repository at the commit we read; the author's words are quoted as theirs.
