Investigate a failed GitHub Actions job and turn its logs into a structured bug issue for Megatron-LM. Collects failing tests, errors, and related pull requests, checks for duplicates, and identifies an assignee from test history.
- Create a bug issue from a failed GitHub Actions job
- Summarize failing tests and relevant log excerpts
- Check for an existing issue covering the same test
NVIDIA/skills☆ 3.7KUpdated Oct 24, 2026
Rebuild Jetson hardware configuration files, kernel images, and modules after source changes. Selects a build mode from the changed repositories or an explicit request, then records the outputs in a manifest for a separate deployment step.
- Rebuilding hardware configuration after a change
- Rebuilding the kernel and modules
- Resuming an interrupted build
NVIDIA/skills☆ 3.7KUpdated Oct 24, 2026
Check Jetson video encoding and decoding capabilities by comparing live API reports, verified NVIDIA samples, and official documentation. Distinguish documented product support from observed behavior for specific formats, bit depths, dimensions, and tested operations.
- Checking support for video formats and bit depths
- Comparing API reports with official product documentation
- Verifying encoding and decoding under specific conditions
NVIDIA/skills☆ 3.7KUpdated Oct 24, 2026
Run a warehouse robot demo controlled by Mission Control and displayed in a local Isaac Sim window. Check startup readiness, replay a fixed route, and verify mission completion, robot movement, and system health.
- Demonstrating Mission Control with a simulated warehouse robot
- Replaying a fixed robot route
- Investigating startup or mission failures
NVIDIA/skills☆ 3.7KUpdated Oct 24, 2026
Edit an existing Isaac for Healthcare simulation scene, including object placement, cameras, and task success rules. Review changes in a running simulator, then save the confirmed scene to its source files when explicitly requested.
- Adjusting object and robot placement in an existing scene
- Adding a camera and checking its view
- Updating task instructions and success rules
NVIDIA/skills☆ 3.7KUpdated Oct 24, 2026
Build and troubleshoot NVMe-oF storage targets that expose local disks over a network using DOCA STA. Guides device support checks, I/O queue sizing, and error diagnosis on BlueField or ConnectX hardware.
- Build an accelerated NVMe-oF storage target
- Check device support and I/O queue limits
- Investigate connection timeouts over RoCE
NVIDIA/skills☆ 3.7KUpdated Oct 24, 2026
Derive HSB Sensor RX packetizer settings from sensor data widths, stream counts, and rearrangement needs. Produce SystemVerilog configuration lines or a YAML handoff, with explanations of how the chosen values shape data handling.
- Choose settings for splitting sensor data into virtual streams
- Configure data rearrangement or replication
- Prepare packetizer values for hsb-ip-def
NVIDIA/skills☆ 3.7KUpdated Oct 24, 2026
Read hardware counters, values describing device activity, from BlueField or ConnectX in a DOCA application. The skill guides supported-counter checks, reader setup, sampling, and error handling across six domains, including PCC, DPA, PHY, and PCI.
- Reading PCC counters from a BlueField application
- Sampling DPA counters
- Adding PHY or PCI counter reads to an application
NVIDIA/skills☆ 3.7KUpdated Oct 24, 2026
Create a minimal Isaac for Healthcare Workflow with a ground plane, dome light, and one idle mode. It generates five scaffold files and runs static checks without launching Isaac Sim, providing a starting point for later scene authoring.
- Starting a blank healthcare robotics Workflow
- Previewing scaffold files before creation
- Checking a new scaffold without launching Isaac Sim
NVIDIA/skills☆ 3.7KUpdated Oct 24, 2026
Find and run Holoscan Sensor Bridge example applications that match a connected development kit, board, software version, and sensors. Monitor timed runs, investigate failures, suggest code or environment changes, and summarize the session.
- Find example apps compatible with connected sensors
- Run a sample application for 60 seconds
- Analyze failures and retry an application
NVIDIA/skills☆ 3.7KUpdated Oct 24, 2026
Assess whether a DOCA application needs low-level communication controls beyond its higher-level APIs. For those cases, guide queue configuration, libibverbs migration, and error diagnosis using installed headers, version checks, and device capability queries.
- Decide whether higher-level DOCA APIs are sufficient
- Configure a queue attribute unavailable through the task API
- Port libibverbs code into DOCA Core
NVIDIA/skills☆ 3.7KUpdated Oct 24, 2026
Integrate a machine learning weather model into Earth2Studio for forecasts that advance through time. The skill guides wrapper implementation, coordinate handling, model loading, tests, dependency declarations, and documentation updates using a reference inference script.
- Add an external weather model to Earth2Studio
- Align model inputs, outputs, and coordinates
- Test forecast steps and sequences
NVIDIA/skills☆ 3.7KUpdated Oct 24, 2026
Helps developers offload remote memory operations from HPC, UCX, or MPI applications to a BlueField DPU, a processor for infrastructure tasks. Covers host-side DOCA UROM setup, supported-operation discovery, completion handling, and error investigation.
- Offloading MPI communication work to a BlueField DPU
- Checking supported operations through plugin discovery
- Investigating errors from host-side UROM calls
NVIDIA/skills☆ 3.7KUpdated Oct 24, 2026
Generate, validate, compare, and explain HSB IP configuration files in HOLOLINK_def.svh. Use bundled scripts to check configuration macros, report errors and warnings, and identify meaningful changes between files while explaining how settings affect ports.
- Generate a configuration file for a new HSB board
- Review errors and warnings in an existing configuration
- Understand macro constraints and their effects on ports
NVIDIA/skills☆ 3.7KUpdated Oct 24, 2026
Check whether a requested Holoscan SDK 5.x release supports Conda installation. When packages are unavailable or unconfirmed, find an alternative installation method that preserves the requested SDK version and C++ or Python requirements.
- Check Conda support for Holoscan SDK 5.x
- Choose an alternative way to install SDK 5.0
- Find an installation method for C++ or Python
NVIDIA/skills☆ 3.7KUpdated Oct 24, 2026
Generate a SystemVerilog top-level scaffold from a validated HOLOLINK_def.svh file. The output declares HSB signals, instantiates HOLOLINK_top, and leaves integration TODOs, using current IP source or bundled references for supported revisions.
- Generating an FPGA_top.sv scaffold from validated definitions
- Creating a matching wrapper after hsb-ip-def validation
- Explaining signal groups controlled by macros
NVIDIA/skills☆ 3.7KUpdated Oct 24, 2026
Install Holoscan SDK 5.x Python packages in an isolated environment after checking platform and CUDA 13 requirements. Choose the full stack or matching components, then verify dependencies, imports, and a release-matched example pipeline.
- Install the full Holoscan SDK 5.x Python stack
- Set up Core with selected optional components
- Check host compatibility before installation
NVIDIA/skills☆ 3.7KUpdated Oct 24, 2026
Implement structured telemetry, such as counters and events, from DOCA applications to external consumers. Guides publishing API selection, schema and source setup, build and test workflows, and debugging lifecycle errors or reports that never reach the receiver.
- Publish packet counters and bandwidth gauges from a DOCA application
- Define event fields and initialize schemas and sources
- Investigate DOCA_ERROR_BAD_STATE during event reporting
NVIDIA/skills☆ 3.7KUpdated Oct 24, 2026
Build Holoscan SDK 5.x from a selected tag, commit, or development branch using its run wrapper. Check prerequisites, choose build options, preserve existing changes, and verify the result with example tests for SDK development or Early Access features.
- Building unpublished Early Access features
- Building and testing a modified SDK checkout
- Creating debug or C++-only builds
NVIDIA/skills☆ 3.7KUpdated Oct 24, 2026
Implement memory copies with DOCA DMA, which transfers data using the device rather than the host CPU. Get guidance on initialization, buffer limits, memory permissions, task completion, and diagnosing errors against the locally installed SDK.
- Implement a host-to-BlueField memory copy
- Check buffer size limits and scatter-gather support
- Investigate permission errors or missing task completion
NVIDIA/skills☆ 3.7KUpdated Oct 24, 2026
Build packet steering rules, which direct network traffic by matching its properties, with DOCA Flow on NVIDIA NICs or DPUs. Adapt shipped samples, validate rules against installed APIs and hardware capabilities, and investigate counters or runtime errors.
- Route traffic matching address and port conditions
- Bring up Flow ports
- Validate rules before programming hardware
NVIDIA/skills☆ 3.7KUpdated Oct 24, 2026
Implement and troubleshoot storage recovery with DOCA Erasure Coding. The skill guides parity generation, lost-block recovery, and incremental parity updates using Reed-Solomon coding, while checking device capabilities, block limits, and memory permissions.
- Recover missing data or parity blocks
- Update parity after one data block changes
- Check device support for a storage layout
NVIDIA/skills☆ 3.7KUpdated Oct 24, 2026
Develop DOCA Rivermax applications that receive timing-sensitive video, audio, and other data over IP networks. Get guidance on receive stream setup, device capability checks, and troubleshooting errors or missing receive events.
- Set up an SMPTE ST 2110 receive stream
- Investigate a stream that receives no frames
- Check supported clock and packet placement options
NVIDIA/skills☆ 3.7KUpdated Oct 24, 2026
Configure DOCA GPI so CUDA kernels can initiate RDMA, direct remote memory access, without host CPU mediation. The skill guides channel setup, memory attachment, GPU handle handoff, library selection, and diagnosis of GPI errors.
- Choose between GPI and GPUNetIO for GPU-driven communication
- Configure GPI channels and remote endpoint connections
- Pass a GPI channel handle to a CUDA kernel
NVIDIA/skills☆ 3.7KUpdated Oct 24, 2026