NVIDIA Introduces DOCA AI Agent Skills on GitHub to Accelerate BlueField DPU Development

NVIDIA Introduces DOCA AI Agent Skills on GitHub to Accelerate BlueField DPU Development NVIDIA has released DOCA AI agent skills on GitHub, providing general-purpose AI agents with machine-readable domain specifications to improve their accuracy and efficiency when developing on NVIDIA BlueField DPUs. DOCA agent skills use an open format centered around SKILL.md files covering API signatures, hardware capability rules, and build constraints across DOCA libraries like Flow, GPUNetIO, and PCC. In a benchmark of 65 prompts, agents without skills satisfied only 19% of checklist requirements and frequently invented APIs or ignored hardware checks, whereas agents equipped with skills met 100%. A side-by-side RDMA test on BlueField-3 showed that an agent with DOCA skills required 73% fewer lines of code and 46% fewer hardware commands to complete the task. Addressing Domain Gaps in Infrastructure AI General-purpose AI agents lack domain-specific awareness of NVIDIA's DOCA platform and BlueField data processing units (DPUs), resulting in pattern-matched guesswork. In testing across 65 developer prompts, unassisted agents made consistent errors, such as misusing APIs and flags in 59 instances, skipping hardware capability verification in 46, and guessing versions in 30. Overall, agents without skills met only 19% of graded checklist items, compared to a 100% success rate across all prompts when equipped with DOCA skills. Core Development Scenarios and Safeguards The skills provide structured contracts that assist agents across key scenarios, including utilizing valid API calls for components like DOCA Comch, verifying hardware topologies before generating code, and providing correct linker flags through tools like pkg-config. For high-risk firmware tasks such as mlxconfig modifications on live hardware, the skills instruct agents to apply preflight checks, require out-of-band paths, configure rollback plans, and account for cold power-cycle requirements. Efficiency Gains and Availability In a direct comparison evaluating RDMA traffic generation on BlueField-3, an agent using DOCA skills generated 189 lines of handwritten code compared to 695 lines from an unassisted agent, while cutting hardware commands from 37 down to 20. The DOCA AI agent skills, covering libraries such as Flow, GPUNetIO, and RDMA, are publicly available in the NVIDIA/skills GitHub repository.

NVIDIA推出DOCA AI Agent Skills 加速BlueField DPU应用开发 NVIDIA宣布在GitHub推出DOCA AI agent skills，通过提供真实API签名、硬件能力要求及构建约束等机器可读规范，弥补通用AI Agent在DOCA领域专业知识不足的问题，大幅减少开发调试周期。 DOCA AI agent skills采用开源的SKILL.md规范，覆盖Flow、GPUNetIO、PCC等全套DOCA库。 在65个真实开发者提示词的评测中，未配备技能的Agent仅满足19%的标准清单项目，而配备技能后达到100%。 未配备技能的Agent高频出现错误：API误用占59/65，未验证硬件能力占46/65，工具路由错误占39/65。 实际构建RDMA程序测试中，具备技能的Agent手写代码量减少73%（189行对比695行），硬件命令减少46%（20次对比37次）。 提供真实API与构建支持 通用Agent在处理DOCA任务时常臆造不存在的函数、参数标志与镜像标签。DOCA Agent Skills使Agent能够直接调用真实的API签名和正确的参数顺序。在编译和链接阶段，技能还能指导Agent利用pkg-config等工具快速定位正确的链接器标志，避免因链接错误陷入无效排错。 硬件能力前置验证与安全变更 未配置技能的Agent往往假设硬件支持某功能后直接写代码，导致在实际硬件上报错。配置技能后，Agent会在编码前核查GPU-NIC PCIe拓扑及GPUNetIO等支持情况。在处理mlxconfig级固件参数修改时，技能要求Agent实施预检清单、设定回滚方案，并提示修改需冷重启生效，有效保障生产环境硬件操作安全。

AI agents are becoming a standard part of development workflows, but general-purpose agents weren’t built with specialized infrastructure software such as NVIDIA DOCA in mind. Without domain-specific knowledge, agents may fall back on guesswork. This is an issue in infrastructure development because every correction cycle takes time away from deployment. DOCA is the unified software platform that unlocks the full potential of NVIDIA BlueField data processing units (DPUs) for agentic AI infrastructure. It spans accelerated networking, AI-native storage, in-silicon security , telemetry, and lifecycle management. It’s the development platform for teams building on NVIDIA BlueField. DOCA AI agent skills are now available on GitHub. These skills provide a structured, verified foundation to address where general-purpose AI agents fall short on DOCA development. They provide agents with verified API signatures, hardware capability requirements, and build constraints. This post explains what DOCA AI agent skills are and how they perform across four impactful DOCA development scenarios. It includes a side-by-side demo and how to get started using DOCA agent skills to build faster and ship more stable code on BlueField infrastructure. What are DOCA AI agent skills? DOCA AI agent skills are a standardized way to give AI agents new capabilities and expertise. The lightweight, open format is built around a SKILL.md file containing real API signatures, hardware capability requirements, and build constraints. Together, the skills provide agents with a reasoning and operational framework for DOCA. Skills span the full DOCA library, including Flow, GPUNetIO, PCC, and more. They don’t replace the agent, but give it the domain knowledge to reason like an experienced DOCA developer. Each skill is scoped to a specific DOCA component or workflow. When an agent loads a skill for DOCA Flow, for example, it gets the real function signatures, the correct pkg-config module names, the build-container constraints, and the common failure modes and their mitigations. The skills are not just a summary of documentation, but a machine-readable specification the agent can reason against directly. Why do AI agents need a foundation for DOCA development? When you ask a general-purpose AI agent to set up a DOCA Comch (Comm Channel), configure an RDMA context, or debug a link failure in a DOCA Flow program, the agent is working from pattern-matching across general training data—not from verified DOCA API contracts, hardware capability manifests, or build system specifications. The DOCA library surface is large, rapidly evolving, and hardware-specific in ways that general training data does not capture. There is no machine-readable contract for agents to reason from, no guaranteed stable interface between what the agent knows and what the hardware and software actually support. To quantify the gap, the NVIDIA team ran 65 real DOCA developer prompts comparing agent performance with and without the skills. These prompts ranged from a one-line question to detailed multi-requirement prompts, graded against a required-answer checklist (a specific set of pass/fail criteria for each task). Without the skills, agents consistently made the same mistakes, including: Misuse of APIs and flags : 59/65 prompts Hardware capability not verified : 46/65 prompts Wrong tool routing : 39/65 prompts Skipped smoke tests : 34/65 prompts Guessed versions : 30/65 prompts These are consistent failures. Without skills, agents satisfied only 19% of graded checklist items on real DOCA tasks. With skills, they satisfied 100% across all 65 prompts. Without the skills, every agent mistake becomes a developer debugging session, and a longer path to working code. What are the benefits of DOCA agent skills for developers? DOCA agent skills allow you to equip agents with domain expertise. This means you can build faster, ship more stable code, and deploy with fewer unknowns. Build faster: Skills give the agent verified API calls from the start, so you can focus on writing code, not corrections. Ship more stable code: With skills, the agent checks what the device actually supports before writing a single line, not after running it on real hardware. Deploy with fewer unknowns: The agent applies preflight checks, rollback plans, and cold power-cycle awareness before touching hardware, not after something goes wrong. The combined effect is fewer correction cycles between task and working code. For teams using AI agents on DOCA applications at scale, that reduction compounds across every developer, task, and deployment. Use agent skills to accelerate DOCA development Each of four capabilities are detailed in the following sections. Each capability is illustrated with a real prompt from our 65-prompt evaluation and the benefits of using the skill. Build faster with real APIs What DOCA AI agent skills address: Misuse of APIs: agents inventing functions, flags, and image tags that don’t exist. This was the most frequent failure mode in our evaluation, affecting 59 of 65 prompts. Without skills, agents generate code referencing DOCA functions that don’t exist in the library, flags with incorrect names, and image tags that fail at runtime. As the developer, you end up debugging an error you didn’t introduce (the agent’s invention). Prompt: “I have a host with a BlueField-3 DPU and DOCA installed. I want to set up a DOCA Comch (Comm channel) between a host-side process and a DPU-side agent so they can exchange small control messages (under 4 KiB each, a few times per second).” The AI agent with skills uses only real DOCA Comch API calls: correct argument order, verified flags and lifecycle sequence, nothing invented. When the agent’s first response uses the correct API surface, you can move directly to integration rather than starting with a debugging session. Across 63 prompts that tested API accuracy, the with-skills result was better every time. Verify hardware before writing any code What DOCA AI agent skills address: Agents writing code for features the device doesn’t support. Hardware capability verification was missed in 46 of 65 prompts without skills. The failure pattern is consistent: the agent assumes a capability is present, generates code against it, and the developer discovers the mismatch only when running on real hardware. At that point, the fix requires understanding which capability check was skipped, reworking the code path, and retesting. Prompt: “I have two hosts wired together with a high-speed InfiniBand fabric. Each host has one NVIDIA GPU and one ConnectX NIC, and DOCA is installed on both. I want to measure the latency of an RDMA WRITE work request when the WR (work request) is posted from a CUDA kernel on the GPU (not from a host CPU thread) so I can decide whether to commit my application to the GPUNetIO Verbs interface.” The AI agent with skills checks what the device actually supports before writing any code. It verifies GPU-NIC PCIe topology, confirms GPUNetIO support before committing, and names alternative API surfaces and when each applies. The AI agent with skills validates the preconditions for the answer against the actual hardware in use, before any code is written. Ship code that compiles and links What DOCA AI agent skills address: Agents producing code that won’t compile or link in the DOCA container. Build correctness was tested on 10 of 65 prompts. This is a smaller subset, but when it applies, development cannot proceed. Prompt: “I’m trying to build a DOCA Flow program from one of the shipped samples. The compile step works but the link fails: undefined reference to doca_flow_init. How do I debug this?” The AI agent with skills identifies the undefined reference as a link-time failure, uses the build tool (pkg-config) to look up the correct linker flags for doca-flow. An agent that generates an incorrect linker flag sends the developer down a path that won’t resolve. The with-skills agent immediately names the correct diagnostic approach and the right tool to retrieve the correct flags. Execute changes that affect hardware safely What DOCA AI agent skills address: Agents applying firmware-level changes on live hardware without preflight checks, rollback plans, or knowledge of cold power-cycle requirements. The more complex the task, the more the skills matter. In our 65-prompt evaluation, the largest performance gap appeared in the highest-complexity scenario: firmware-level changes on live hardware. Prompt: “I have a production BlueField-3 DPU running a DOCA workload. I loaded the DOCA AI agent skill for my service, and the next step is to write an mlxconfig-class firmware-level parameter.” The AI agent with skills applies the full discipline required for firmware-level changes to meet every requirement. This includes preflight inventory, assuming an out-of-band (OOB) path as a precondition, an explicit maintenance window, a rollback plan, and notes that mlxconfig-class writes take effect only on a cold power cycle, not warm reboot. The with-skills agent met every requirement. The without-skills agent met none. On a live production DPU, none of these misses are recoverable debugging steps. The skills carry the knowledge of what can go wrong and why, turning a risky firmware change into a disciplined, safe procedure. Across all 65 prompts, AI agents with DOCA skills reached the correct answer every time. This means fewer correction cycles and faster time to working code. The demo in Video 1 puts the efficiency gain in concrete terms. Two agents were given the same task: build a program using NVIDIA DOCA to send real RDMA traffic on BlueField-3. Both succeeded, but the with-skills agent used 73% less handwritten code (189 lines versus 695) and roughly half the hardware commands (20 versus 37, a 46% reduction). That’s the difference between an agent that rediscovers DOCA APIs and build requirements by trial and error, and one that already has them before writing a single line. Video 1. Watch a side-by-side comparison of two AI coding agents tasked with building a Go-based RDMA application using NVIDIA DOCA on a BlueField-3 platform Get started with NVIDIA DOCA AI agent skills DOCA AI agent skills give your agent the domain knowledge to succeed so you can build with more efficiency and accuracy. To get started, visit the NVIDIA/skills GitHub repo. Check out the general skills for a quick onramp and the library-specific skills for DOCA Flow, GPUNetIO, RDMA, and more.