dx-agent-verify

Validates DX-Stream pipeline builds through a six-step verification gate before completion claims.

3|5|Updated Jan 21, 2025
One-click install
npx skills add https://github.com/DEEPX-AI/dx_stream --skill dx-agent-verify-deepx-ai
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: dx-agent-verify
Source: https://github.com/DEEPX-AI/dx_stream/tree/main/.deepx/skills/dx-agent-verify
Command: npx skills add https://github.com/DEEPX-AI/dx_stream --skill dx-agent-verify-deepx-ai

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? AI agents often claim a GStreamer pipeline build is done without actually verifying it, leading to broken pipelines, silent preprocess-id mismatches, and missing files. This Skill enforces an evidence-based verification gate so no completion is claimed without fresh command output. ## Core Features & Use Cases - Six-Step Verification Gate: Checks file existence, Python/shell/JSON syntax, pipeline --help parsing, DX element property validation, GStreamer plugin registration, and portability of the app directory. - Pipeline Property Checks: Validates preprocess-id matching between dxpreprocess and dxinfer, queue placement between DX elements, absolute model paths, DxMsgConv-before-DxMsgBroker ordering, and DxRate presence for RTSP sources. - Completion Report Template: Produces a structured checklist report with evidence for every item before any completion claim is allowed. - Use Case: After an agent generates a DX-Stream object-detection pipeline with run scripts and configs, invoke this Skill to run the full gate and produce a verified completion report before marking the task done. ## Quick Start Ask the agent to run the dx-agent-verify gate on the current pipeline session directory and produce the completion report before claiming the build is finished.

Frequently Asked Questions about dx-agent-verify

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I verify a GStreamer pipeline build before marking it done?▼

Run the six-step verification gate: check required files exist, validate Python/shell/JSON syntax, run pipeline.py --help, validate DX element properties, confirm plugin registration, and test portability. Only claim completion after producing the full report with all items checked.

How do I check preprocess-id matching between dxpreprocess and dxinfer?▼

Extract preprocess-id values from both elements in the run script using grep and compare the sorted lists. A mismatch causes silent inference failures, so the gate fails the build until the IDs align.

Does the verification gate handle RTSP and broker pipelines differently?▼

Yes. RTSP pipelines must include DxRate, and broker pipelines must place DxMsgConv before DxMsgBroker and include broker config files. These conditional checks run only when the corresponding elements are detected in the run script.

Why does pipeline.py --help fail during validation?▼

A failing --help usually indicates import errors or broken argument parsing in the pipeline script. Fix the underlying Python errors before proceeding, since the gate treats this as a blocking failure.

What makes a DX-Stream app directory portable?▼

The app must run when copied outside the suite: no symlinks escaping the directory, no hardcoded suite paths in code, and model paths resolved via $SUITE_ROOT auto-detection. The gate copies the app to a temp directory and re-checks these conditions.