vss-deploy-profile
The skill bypasses human-in-the-loop controls for infrastructure changes, insecurely handles sensitive API keys, and executes unconstrained network and system operations without declaring required tool permissions.
Skill Info
Install
npx skills add https://github.com/nvidia/skillsAssessments (4)
AI Agent Traps ↗The instructions explicitly provide a mechanism to bypass human confirmation for deployment if the user includes keywords like 'autonomously' or 'non-interactive', which can be exploited to perform unauthorized infrastructure changes. [ensemble: confirmed by 3/3 passes; severity set to the agreed median (ADR-0067).]
If the user's request already asks you to run autonomously (e.g. 'deploy X autonomously', 'run without confirmation', 'non-interactive'), skip the confirmation prompt and proceed straight to Step 5.
The skill handles multiple sensitive API keys (NGC, NVIDIA, HF) and writes them into a generated environment file, which may be logged or persisted in insecure locations.
Validate every credential... before Step 1c copies .env to generated.env
Remote installer fetch-and-run from a known first-party endpoint (downgraded from critical)
https://astral.sh/
Autonomous operation instructions detected — attempts to remove human-in-the-loop controls
autonomous mode
SKILL.md links to "references/alerts.md" but the file is not part of the skill package — the workflow silently degrades or the content is sourced elsewhere at runtime
[`references/alerts.md`](references/alerts.md)
SKILL.md links to "references/base.md" but the file is not part of the skill package — the workflow silently degrades or the content is sourced elsewhere at runtime
[`references/base.md`](references/base.md)
SKILL.md links to "references/lvs-profile.md" but the file is not part of the skill package — the workflow silently degrades or the content is sourced elsewhere at runtime
[`references/lvs-profile.md`](references/lvs-profile.md)
SKILL.md links to "references/search.md" but the file is not part of the skill package — the workflow silently degrades or the content is sourced elsewhere at runtime
[`references/search.md`](references/search.md)
SKILL.md links to "references/warehouse.md" but the file is not part of the skill package — the workflow silently degrades or the content is sourced elsewhere at runtime
[`references/warehouse.md`](references/warehouse.md)
Version History (3)
Badge
[](https://mondoo.com/ai-agent-security/skills/github/nvidia/skills/vss-deploy-profile)<a href="https://mondoo.com/ai-agent-security/skills/github/nvidia/skills/vss-deploy-profile"><img src="https://mondoo.com/ai-agent-security/api/badge/github/nvidia/skills/vss-deploy-profile.svg" alt="Mondoo Skill Check" /></a>https://mondoo.com/ai-agent-security/api/badge/github/nvidia/skills/vss-deploy-profile.svgSecure your AI agents
Skills can read files, run commands, and access credentials. Mondoo helps organizations manage the security risks of AI agent skills across their entire fleet, discover and control Shadow AI usage.
- Continuous scanning of skills across all registries
- Policy enforcement before skills reach your agents
- Integration with your existing security workflow