Build Applications on NVIDIA BlueField Faster with NVIDIA DOCA Agent Skills
- Source
- developer.nvidia.com
- Author
- Tanya Lenz
- Date
Context
NVIDIA's DOCA skills give coding assistants concrete specifications for its BlueField infrastructure software. The files supply actual API signatures, hardware requirements and build constraints, addressing errors such as inventing a function or assuming a device supports a feature. The technical article explains how an agent can consult this material before proposing code.
In NVIDIA's comparison using 65 developer prompts, agents satisfied 19% of graded checklist requirements without the skills and 100% with them. That is evidence of improved answer compliance on the company's selected tasks. The mechanism is explicit domain knowledge supplied at task time; the measured percentage is not a general production-code reliability rate.
- Each DOCA is a SKILL.md scoped to one component, holding real function signatures, pkg-config module names, build-container constraints and known failure modes. NVIDIA frames it as a machine-readable spec, not a docs summary.
- Without skills, the most common failures in NVIDIA's 65-prompt were API and flag misuse (59 prompts), unverified hardware capability (46), wrong tool routing (39), skipped smoke tests (34) and guessed versions (30).
- For a firmware-level mlxconfig change on a production BlueField-3, NVIDIA says the with-skills agent covered preflight inventory, an out-of-band path, a maintenance window and rollback, and noted the write needs a cold power cycle, not a warm reboot.
- In NVIDIA's side-by-side demo building a Go RDMA app on BlueField-3, both agents succeeded, but the one with skills wrote 189 lines of handwritten code versus 695 and ran 20 hardware commands versus 37.
- AI agent — An AI system that doesn't just answer once but works toward a goal in a loop — taking actions, reading the results, and deciding what to do next.
- eval — A repeatable test for AI quality — a set of tasks plus scoring — used the way software teams use test suites, because model output is too variable to judge by eyeballing.
- agent skill — A reusable instruction file that teaches an agent how to do one job well — the procedure, the tools, and what counts as done.
articleScaling NVIDIA TensorRT Performance Optimization Beyond a Single Build Route with the Global Performance Tuner, Part 2
articleAccelerating Mamba-2 with High-Performance Forward Pass Kernels in NVIDIA CUDA Tile
articleScale Agentic Coding Workflows with Meta’s Muse Spark Open on NVIDIA GB300 NVL72
articleOptimizing Agent Reliability and Tool Calling Efficiency for Chip Design Using NVIDIA NeMo Gym
Checking sign-in…
Loading comments…