skilld
All skills
nvidia avatar

/deepstream-sop

@4cb1092
by NVIDIA Corporationnvidia/skills3.5k stars
424

Use this skill when building, deploying, evaluating, debugging, or measuring latency for the DeepStream SOP Inference Microservice — a GPU-accelerated FastAPI service that detects whether operators perform assembly-line steps in order via event boundary detection (GEBD) plus VLM classification. Trigger even if the user does not name it: verify operator step sequence, detect missing or out-of-order SOP steps, score factory/work-cell video for procedure compliance, run VLM-based SOP checking on industrial cameras, or call /v1/chat/completions with a file, RTSP, or Basler camera. Also trigger for its internals: SOPVideoProcessor, DeepStream GEBD model (e.g. DDM) via Triton CAPI, nvds_custom_postprocess, Cosmos Reason 1/2 vLLM, SSE streaming, Kafka NvProto/JSON output, Basler/Pylon camera + emulation, Docker compose, chunk-level latency. Do NOT trigger for generic DeepStream pipelines, object detection/tracking, NIM imports, or video summarization.

Ask your Agent

Use this Skill: https://skilld.dev/gh/nvidia/skills/deepstream-sop

This session only. Nothing lands on disk.

referenceseval_sop_prompt.md

≈220 tokens on demand. Your agent reads this file only when SKILL.md points to it.

SOP Inference Microservice Evaluation Prompt

Evaluate @ds_sop_microservice from scratch using the deepstream-sop skill.

Read first:

  • deepstream-sop/SKILL.md
  • deepstream-sop/references/skill_12_evaluation_workflow.md

Use the environment settings provided by the user for:

  • DDM_MODEL_PATH
  • VLLM_MODEL_PATH
  • MODEL_ROOT_DIR
  • HOST_CACHE
  • MEDIA_STORAGE_DIR
  • ACTION_CONFIG_PATH
  • VLM_PROMPT_PATH
  • VLM_FPS
  • TEST_VIDEO_PATH
  • optional Basler camera serial and live-camera chunk settings

Required outcome:

  1. Run static validation before building.
  2. Build the Docker image.
  3. Launch the service with deploy/.env.
  4. Run the API unit tests.
  5. Run file-video API evaluation.
  6. Run requested latency, camera, or Kafka evaluation.
  7. Fix issues found during evaluation.
  8. Record commands, results, timings, and fixes in a Markdown report.

Source: SKILL.md on GitHub

Third-party checks

2 warnings27d3 checks · Risk MEDIUM
  • Gen Agent Trust Hub27d

    This skill provides a framework for building an industrial SOP monitoring microservice. It uses multi-stage processing, including Vision-Language Models and custom logic. Security observations include dynamic code generation based on configuration files, execution of build scripts via subprocess, and reliance on external research repositories for model weights and code.

  • Socket27d

    2 alerts: gptSecurity, gptAnomaly

  • Snyk27d

    Risk: LOW · No issues

Provenance

Signed by skilld at 4cb1092. This ties the file your Agent reads to that commit on GitHub. It does not review the instructions.

Last checked against GitHub yesterday.

Activeupdated 3 months ago

Capability

owner
windy@nvidia.com
service
deepstream-sop
version
1.0.0
Other metadata
reviewed
2026-04-08
metadata
{
  "author": "Wind Yuan <windy@nvidia.com>",
  "tags": [
    "deepstream",
    "sop",
    "vlm",
    "triton",
    "gpu"
  ],
  "languages": [
    "python"
  ],
  "frameworks": [
    "deepstream",
    "triton",
    "fastapi"
  ],
  "domain": "video-analytics"
}

README badge

README badge for nvidia/skills/deepstream-sop