What does this skill do, and when should you use it?
This skill is for VSS users who need to run the vss-behavior-analytics service on its own. It routes operators through entrypoint selection, configuration sourcing, and optional calibration, with a Docker Compose bring-up example. It requires a VSS repository checkout, NGC credentials, specified Docker versions, and suitable host resources. When a broker is reachable, configuration and calibration can be updated at runtime through mdx-notification without redeployment.
It checks prerequisites including the repository path, NGC API key, Docker versions, and an optional message broker; guides selection of analytics 2D, 3D, mv3dt, dev_example, or fusion_search entrypoints; uses profile-shipped or custom configuration and calibration files; starts the vss-behavior-analytics-base container with Docker Compose; describes probing /docs or /health and handling NGC authentication and GPU-memory failures; and, with a reachable broker, explains publishing dynamic configuration and calibration messages, persistence, worker application, and acknowledgements.
- A VSS operator wants behavior analytics running without deploying the complete warehouse stack.
- A user needs to select an analytics 2D, 3D, mv3dt, dev_example, or fusion_search entrypoint.
- A deployment must use custom behavior-analytics configuration or calibration JSON.
- A running service needs configuration or calibration changes without a restart, and a broker is available.
How do you install this skill?
- Execution changes Compose deployment state and pulls an image from nvcr.io; confirm the target host, image source, permissions, and rollback procedure before deployment.
- Without a broker, the container enters a restart loop under restart: always rather than remaining stably idle.
- The NGC key is sensitive; inject it through the current session or a managed secret and avoid storing it in the repository, shell profile, or logs.
- Dynamic calibration has no ACK and failures are primarily observable through container logs; both dynamic protocols have strict validation and may drop malformed messages.
- No Chinese documentation or mainland-China reachability evidence is provided; verify network access to NGC and related image services in advance.
- Shell / CLI
- Network access
- Local filesystem
Docker Engine 28.3.3Docker Compose plugin v2.39.1+NGC_CLI_API_KEYVSS repository checkout
Install the NVIDIA skills collection with npx skills add nvidia/skills. To install this skill directly, use npx skills add nvidia/skills --skill vss-setup-behavior-analytics --yes. The CLI prompts for the skill and installation destination; the source does not specify a fixed destination directory.
tmp="$(mktemp -d)"
git clone --depth 1 https://github.com/NVIDIA/skills.git "$tmp"
mkdir -p ~/.claude/skills
cp -R "$tmp/skills/vss-setup-behavior-analytics" ~/.claude/skills/
rm -rf "$tmp"Generated from the source repository and skill path; it copies only this skill's folder. If the author's install steps above differ, follow those first. To scope it to one project, replace ~/.claude/skills with that project's .claude/skills.
How do you use this skill?
Once installed, send your agent any of these to trigger it:
- deploy behavior analytics standalone
- run behavior-analytics without the full stack.
Ask a compatible agent to “deploy behavior analytics standalone” or “run behavior-analytics without the full stack.” Minimal bring-up: change to <repo>/deploy/docker, set VSS_APPS_DIR=$(pwd), then run docker compose -f services/analytics/behavior-analytics/compose.yml up -d vss-behavior-analytics-base. The source places the complete entrypoint, configuration, calibration, and dynamic-update procedures in referenced documentation.
What are this skill's strengths and limitations?
- Focused specifically on standalone vss-behavior-analytics deployment.
- Supports profile-shipped and custom configuration and calibration.
- Enables runtime configuration and calibration updates without restarting when a broker is available.
- Includes troubleshooting directions for connectivity, NGC authentication, and GPU memory failures.
- Requires the matching VSS profile or microservice to be deployed and reachable.
- Requires Docker, NGC credentials, the repository layout, and sufficient GPU resources for the selected profile.
- Without a broker, the container may enter a restart cycle and runtime updates are unavailable.
- The source provides a Tier-3 evaluation command but no evaluation results or benchmark data.
How does this skill compare with similar options?
Side by side with related skills; every score comes from the same FSRS standard.
| Skill | FS score | Stars | Last updated | License |
|---|---|---|---|---|
| VSS Standalone Behavior Analytics this page ✓ NVIDIA · Official | 51 · Use with care | ★ 3.6k | 13d ago | Apache-2.0 |
| VSS Standalone Video Analytics API ✓ NVIDIA · Official | 54 · Use with care | ★ 3.6k | 13d ago | Apache-2.0 |
| VSS Video Embedding Deployment ✓ NVIDIA · Official | 57 · Use with care | ★ 3.6k | 13d ago | Apache-2.0 |
| AutoMagicCalib Calibration Stack Launcher ✓ NVIDIA · Official | 54 · Use with care | ★ 3.6k | 13d ago | Apache-2.0 |
| RT-VLM Dense Captioning Deployment ✓ NVIDIA · Official | 45 · Not recommended | ★ 3.6k | 13d ago | Apache-2.0 |
Unlike a full warehouse deployment, this skill launches only the vss-behavior-analytics container. Requests for the complete UI, agent, and perception stack should be routed to vss-deploy-profile.
How did FollowSkills review this skill?
The documentation addresses NGC-key handling, avoiding plaintext command-line exposure, external registry login, and container mounts, while scoping the service to standalone deployment. However, deployment changes Compose state and pulls an external image; explicit pre-execution confirmation, least-privilege requirements, rollback, and complete sensitive-data-flow disclosure are missing. The supplied license metadata is also NOASSERTION, so points are deducted.
Entrypoint, configuration, calibration, broker behavior, troubleshooting, and failure feedback are described in useful detail. Still, this is a static review with no verifiable CI plus committed key-path test execution evidence, and the workflow depends on brittle external paths and exact runtime versions, so the score is conservatively limited.
The name, use cases, routing boundary against full-stack deployment, entrypoint choices, and broker conditions are reasonably clear. Chinese-language support and mainland-China reachability are not addressed; core image access and NGC authentication depend on overseas services, and negative-trigger and environment-variant coverage is limited, so points are deducted.
The material is layered into overview, referenced workflows, troubleshooting, dynamic protocols, limitations, and testing guidance, and declares version 3.2.0, author, license, and constraints. However, references are deeply nested, the author format is incomplete, maintenance ownership and changelog/update procedures are unclear, and the benchmark reports quality and efficiency findings, so points are deducted.
The skill can guide entrypoint selection, config/calibration mounting, container startup, verification, and runtime updates, making the core task plausibly usable. However, the actual Compose file and complete dependency environment are outside the selected skill evidence, NGC, broker, and calibration conditions materially affect success, and the single-task evaluation reports only 50% Codex correctness and effectiveness, so the score remains moderate.
Evaluation manifests, a benchmark report, test-path references, and detailed wire contracts provide some auditability. Coverage is thin: there is only one evaluation task, no negative task, the report is not independent reproduction, and no verifiable CI evidence covers the skill's key paths, so points are deducted.
Open a dimension to read why it scored that way
Evidence confidence：Low — Mostly static review, author material or a limited demo; useful for discovery, not high-risk decisions.
See the full review method →