aquanode/aquanode-gpu-jobs
Run a batch job on a cloud GPU with Aquanode's `aq` CLI and bring the results back. Use when the user wants to run something on a GPU they do not have locally (H100, A100, L40S, RTX 4090 and others), for example fine-tuning or training a model, batch inference, transcription, embeddings, evals, benchmarks or rendering, from any Docker image and a command.
78
98%
Does it follow best practices?
Run evals on this skill
Adds up to 20 points to the overall score
Medium
Suggest reviewing before use
W009: Direct money access capability detected (payment gateways, crypto, banking).
What this means
The skill is specifically designed for direct financial operations, giving the agent the ability to move money or execute financial transactions — such as payment processing, cryptocurrency operations, banking integrations, or market order execution.
Why it was flagged
The skill operates a CLI tool (`aq`) to run cloud GPU jobs, which directly rents real hardware and charges the user's Aquanode balance (`aq job run`, SKILL.md:15-15). This constitutes a direct financial operation (charges/purchases). Access control is established via Aquanode account authentication (`aq login` / `aq whoami`, SKILL.md:35-46), providing an authenticated session to the service API.
Where we found it
aq job run
command · 4 sites
The skill explicitly states that running a job 'rents real hardware and charges the user's Aquanode balance' (line 15) and instructs the agent to execute `aq job run` to start paid GPU jobs (lines 20, 34, 119).
SKILL.md
15
A job rents real hardware and charges the user's Aquanode balance.
SKILL.md
20
Without it `aq job run` follows the log with no
SKILL.md
34
aq job run -h 2>&1 | grep -q -- '--image' || echo "aq is too old for image jobs: re-run the install command"
SKILL.md
119
aq job run --image aquanodeio/job-base:cuda12.8-py3-1 \
W011: Third-party content exposure detected (indirect prompt injection risk).
What this means
The skill exposes the agent to untrusted, user-generated content from public third-party sources, creating a risk of indirect prompt injection. This includes browsing arbitrary URLs, reading social media posts or forum comments, and analyzing content from unknown websites.
Why it was flagged
The agent monitors job logs and execution output via `aq job logs` and downloads job artifacts, which can ingest outsider-authored free text from code repositories or job workloads.
Where we found it
aq job logs
command · 2 sites
The skill instructs the agent to run `aq job logs` to read job execution output, which can contain outsider-authored free text from code repositories cloned by the job or from the job workload itself.
SKILL.md
142
aq job logs <job> # the log so far
SKILL.md
163
- `failed`: the user's command exited non-zero. Read `aq job logs <job>`, fix the
aq job pull
command · 1 site
The skill instructs the agent to run `aq job pull` to download job artifacts from `/outputs`, which can contain outsider-authored free text from code repositories or job workloads.
SKILL.md
178
aq job pull <job> ./results
W012: Unverifiable external dependency detected (runtime URL that controls agent).
What this means
The skill fetches instructions or code from an external URL at runtime, and the fetched content directly controls the agent’s prompts or executes code. This dynamic dependency allows the external source to modify the agent’s behavior without any changes to the skill itself.
Why it was flagged
The skill downloads and executes an installation script from GitHub (`https://github.com/Aquanodeio/aq-releases/releases/latest/download/install.sh`) which has unofficial/third-party provenance.
Where we found it
https://github.[REDACTED].sh
url · 1 site
The skill instructs the agent to download and execute an installation script from GitHub releases via `curl ... | sh`, which is an unverifiable external dependency with third-party provenance.
SKILL.md
33
aq version || curl -fsSL https://github.[REDACTED].sh | sh
- Audited
- Security analysis