Doca Pipeline Language Developer Tools

NVIDIAOfficialCreated by NVIDIA

Skill
Tool integration
Software development
Apache-2.0
Medium risk

From the author

Use this skill when the operator is driving the DOCA Pipeline Language (DPL) Developer Tools to inspect or debug DPL programs on a BlueField DPU — DPL Nspect for runtime state such as tables, keys, and counters, DPL Debugger for packet flow, DPL Admin for dynamic configuration changes in the DPL Runtime Daemon, P4 Runtime Shell as a P4Runtime control plane manager, or DPL Runtime Controller SDK for custom P4Runtime applications and DPL shared memory APIs. Trigger even when the user does not explicitly mention "DPL Developer Tools" — implicit phrasings include "inspect the DPL tables on the DPU", "why can't I start a DPL debug session", "change a dpl_rtd configuration item without restarting", "build my own P4Runtime app for DPL", or "visualize packet flow through my DPL program". Refuse and route elsewhere: authoring, compiling and deploying DPL programs to the DOCA Pipeline Language Service Guide, DOCA install/bring-up to doca-setup, and hardware or firmware state changes to doca-hardware-safety.

— the description by NVIDIA

AutoGineer summary

Written by AutoGineer from the skill’s own files, not by the author

This skill assists operators in utilizing the DOCA Pipeline Language (DPL) Developer Tools suite, which includes DPL Nspect, DPL Debugger, DPL Admin, P4 Runtime Shell, and the DPL Runtime Controller SDK. It is designed for tasks involving the inspection of runtime state, visualization of packet flows, dynamic configuration of the DPL Runtime Daemon (dpl_rtd), and development of custom P4Runtime applications. The skill assumes a pre-configured environment where the DPL Runtime Service container is active on a BlueField DPU. It explicitly excludes tasks related to DPL program authoring, compilation, deployment, initial DOCA setup, or hardware/firmware modifications.

What it does

Inspects DPL runtime state including tables, keys, and counters using DPL Nspect.
Visualizes packet flow through DPL programs using DPL Debugger.
Modifies configuration items dynamically in the DPL Runtime Daemon via DPL Admin.
Supports the development of custom P4Runtime applications using the DPL Runtime Controller SDK.
Facilitates control plane management through the P4 Runtime Shell.
Validates prerequisites for debug sessions, such as the use of the -g debug flag.

When to use it

When you need to inspect active DPL runtime tables or counters on a DPU.
When you need to debug or visualize packet paths through a loaded DPL program.
When you need to change dpl_rtd configuration settings without a service restart.
When you are building or integrating custom P4Runtime control-plane applications.
When you need to troubleshoot connectivity between developer tools and the DPL Runtime Daemon.

Advantages

Provides a structured approach to debugging and inspecting DPL programs.
Clearly defines scope boundaries to prevent misconfiguration of hardware or firmware.
Integrates with established DOCA workflows for runtime management.
Ensures consistent use of developer tools across different DPL environments.

Usage examples

“How do I inspect the DPL runtime tables on the DPU?”: The agent provides instructions on using DPL Nspect to list active tables, keys, and counters within the dpl_rtd process.
“I want to visualize packet flow through my DPL program.”: The agent guides you through using DPL Debugger, confirming that your program was compiled with the -g flag and loaded via the P4Runtime controller.
“How do I dynamically change a configuration item in dpl_rtd?”: The agent instructs you on using DPL Admin to modify runtime configuration items without requiring a full service restart.
“Can I build a custom P4Runtime application for DPL?”: The agent provides guidance on utilizing the DPL Runtime Controller SDK to develop custom control-plane applications.

Install

Download the skill as a ZIP:
https://autogineer.com/agent-skills/nvidia/doca-pipeline-language-developer-tools/download

Unzip it into the cross-client skills folder of your project:

mkdir -p .agents/skills && unzip doca-pipeline-language-developer-tools.zip -d .agents/skills

Documentation by NVIDIA

DOCA Pipeline Language Developer Tools

Scope boundaries

These requests are out of scope: say so, name the skill that owns the request, and stop.

installing DOCA and bringing up the DPL Runtime Service container: doca-setup/SKILL.md. Name doca-setup in the answer and route the whole install/bring-up there; do not walk the install steps from this skill or from the DPL Installation Guide.
DPU mode or firmware state changes: doca-hardware-safety/SKILL.md.
which DOCA version is installed, version compatibility: doca-version/SKILL.md.
authoring, compiling and deploying DPL programs (DOCA Pipeline Language Service Guide): doca-public-knowledge-map/SKILL.md.

Examples

"How do I inspect the DPL runtime tables on the DPU?" — this is in scope. Use DPL Nspect; do not author or compile DPL programs here.
"I want to visualize packet flow through my DPL program." — this is in scope. Use DPL Debugger; ensure the program was compiled with the debug flag -g and loaded into dpl_rtd using the P4Runtime controller.
"How do I dynamically change a configuration item in dpl_rtd?" — this is in scope. Use DPL Admin; inspect current state first.
"Can I build a custom P4Runtime application for DPL?" — this is in scope. Use the DPL Runtime Controller SDK.
"My DPL program is loaded but I cannot start a debug session." — check the debug-flag prerequisite and the P4Runtime load path before assuming the tool is broken.

Quick facts

The authoritative source for the tool names, flags, and paths is the public DOCA Pipeline Language Developer Tools guide for your release; confirm them with the installed tool's help before quoting any of them. As that guide describes it:

The developer tools are a suite for validating and debugging DPL applications: DPL Compiler, P4 Runtime Shell, DPL Nspect, DPL Debugger, DPL Admin, and DPL Runtime Controller SDK.
All tools interact with the DPL Runtime Daemon (dpl_rtd) running on the BlueField DPU.
DPL Compiler converts DPL source code into a binary format that can be loaded by dpl_rtd.
P4 Runtime Shell is an open-source control plane manager implementing the P4Runtime API.
DPL Nspect is a command-line interface for inspecting DPL runtime state, including tables, keys, counters, and more.
DPL Debugger is a graphical interface for visualizing and debugging packet flow through the DPL program.
DPL Admin is a CLI tool for dynamically modifying configuration items in the DPL Runtime Daemon.
DPL Runtime Controller SDK is an SDK for building custom P4Runtime applications or integrating with DPL shared memory APIs.
Prerequisites: the DPL Runtime Service container must be deployed and running on the DPU, the dpl_rtd process must be active, and the DPU must be configured to operate in DPU mode.
Starting a debug session requires the DPL program to be compiled for debugging: compile with the DPL Compiler debug flag -g enabled, then load the compiled binary into dpl_rtd using the P4Runtime controller.
The developer tools are included in the DPL Development Container, which is publicly available on NGC. Installation instructions refer to the DPL Installation Guide.
For instructions on creating, compiling, and deploying DPL programs, refer to the DOCA Pipeline Language Service Guide.
Support and feedback: visit the NVIDIA DOCA Enterprise Networking Forum. Registration with the NVIDIA Developer Program is required to post; registration is not required to view existing discussions. For help with registration, refer to the FAQ.

Instructions

Follow Loading order below. Open the tasks companion for inspect / debug / modify / build / shell. Open the capabilities companion for what the tools inspect and change. Do not open every companion file.

Where to start: This is a tool skill for invoking the DOCA Pipeline Language Developer Tools — the suite for inspecting and debugging DPL programs on BlueField DPUs. Open the tasks companion and start at ## inspect for DPL Nspect runtime-state inspection, ## debug for DPL Debugger packet-flow visualization, ## modify for DPL Admin dynamic configuration changes, ## build for DPL Runtime Controller SDK custom P4Runtime applications, and ## shell for P4 Runtime Shell control-plane use. Open the capabilities companion when the question is what state can the tools inspect, what do they change, and what prerequisites gate their use. If the user has not installed DOCA yet, route to doca-setup/SKILL.md first. If the user is asking about DPL program authoring, compiling, or deploying rather than inspecting or debugging a loaded program, route to the DOCA Pipeline Language Service Guide, not this skill. Any step that changes hardware or firmware state routes through doca-hardware-safety/SKILL.md. Version questions route through doca-version/SKILL.md; installs route through doca-setup/SKILL.md.

How to use this skill. Read this file, then search the companion file for the ## <name> heading the question needs and read that whole section. Answer completely from it, quoting only its symbols and commands. Do not browse docs; use the entry points in doca-public-knowledge-map. Give environment checks as commands for the user's DOCA host, quoting only commands documented in the public guide or installed tool help. For out-of-scope requests, name the skill that owns them and stop.

Example questions this skill answers well

The CLASSES of DPL developer-tools questions this skill is built to answer, each with one worked example. The class is the load-bearing piece; the worked example is one instance.

"What tools are available for inspecting and debugging DPL programs?" — worked example: "I have a DPL program loaded on the DPU and need to see its tables, keys, counters, or packet flow". Answered by the tool inventory in the capabilities companion + the tasks companion.
"How do I inspect DPL runtime state?" — worked example: "I want to list the tables and counters currently active in dpl_rtd". Answered by the DPL Nspect workflow in the tasks companion.
"How do I debug packet flow through a DPL program?" — worked example: "packets are not matching the expected table entries; I need to visualize the path through the DPL program". Answered by the DPL Debugger prerequisites (debug flag -g, P4Runtime load) in the capabilities companion + the tasks companion.
"How do I dynamically change configuration items in dpl_rtd?" — worked example: "I need to update a configuration item without restarting the DPL Runtime Daemon". Answered by the DPL Admin workflow in the tasks companion.
"How do I build a custom P4Runtime application or integrate with DPL shared memory APIs?" — worked example: "I want to write my own control-plane application for DPL instead of using the P4 Runtime Shell". Answered by the DPL Runtime Controller SDK workflow in the tasks companion.
"What are the prerequisites for using the developer tools?" — worked example: "the tools cannot see dpl_rtd; is the DPL Runtime Service container running and is the DPU in DPU mode?". Answered by the prerequisites in the capabilities companion + the tasks companion.

When to load this skill

Load this skill when the user is — or the agent needs to — drive the DOCA Pipeline Language Developer Tools on a real BlueField DPU with the DPL Runtime Service running. Concretely:

Inspecting DPL runtime state (tables, keys, counters) with DPL Nspect.
Visualizing or debugging packet flow through a DPL program with DPL Debugger.
Dynamically modifying configuration items in dpl_rtd with DPL Admin.
Building custom P4Runtime applications or integrating with DPL shared memory APIs using the DPL Runtime Controller SDK.
Using the P4 Runtime Shell as a P4Runtime control plane manager.
Preparing a debug session: ensuring the DPL program was compiled with the debug flag -g and loaded into dpl_rtd using the P4Runtime controller.
Diagnosing why the developer tools cannot interact with dpl_rtd (DPL Runtime Service container not running, dpl_rtd not active, DPU not in DPU mode).

Do not load this skill for general DOCA orientation, DPL program authoring/compilation/deployment, DOCA install, or hardware/firmware state changes. For those, route to doca-public-knowledge-map/SKILL.md, the DOCA Pipeline Language Service Guide, doca-setup/SKILL.md, or doca-hardware-safety/SKILL.md.

What this skill provides

This is a thin loader. Substantive material lives in two companion files:

the capabilities companion — what the developer tools inspect and change: the tool inventory (DPL Compiler, P4 Runtime Shell, DPL Nspect, DPL Debugger, DPL Admin, DPL Runtime Controller SDK), the read-only vs state-changing operation split (Nspect and Debugger read-only; Admin modifies; Compiler and SDK build), the prerequisites (DPL Runtime Service container, dpl_rtd active, DPU mode, debug flag -g, P4Runtime load), availability via the DPL Development Container on NGC, support/feedback via the NVIDIA DOCA Enterprise Networking Forum, and the high-stakes safety policy that dynamic configuration changes through DPL Admin affect the running DPL Runtime Daemon and must be gated on a clean inspection.
the tasks companion — step-by-step workflows for the in-scope task verbs: inspect (DPL Nspect runtime state), debug (DPL Debugger packet flow, including the debug-flag prerequisite), modify (DPL Admin dynamic configuration), build (DPL Runtime Controller SDK custom P4Runtime applications / shared memory integration), shell (P4 Runtime Shell control plane), plus a Deferred task verbs block and a Command appendix that honors the bundle's doca-structured-tools-contract/SKILL.md preamble.

The skill assumes a BlueField DPU where the DPL Runtime Service container is already deployed and running, dpl_rtd is active, and the DPU is configured to operate in DPU mode.

What this skill deliberately does not ship

This skill is agent guidance, not a samples or scripts bundle. To keep the boundary clean, it deliberately does not contain — and pull requests should not add:

Verbatim subcommand names, flag inventories beyond the documented debug flag -g, default paths, port numbers, or output column names. The public DOCA Pipeline Language Developer Tools guide on docs.nvidia.com and the installed tool help on the user's version are the joint source of truth; copying them here pins the skill to one release and silently rots when the tools evolve. The skill routes the agent at those sources instead.
Pre-baked example output. Output is install-, version-, and deployment-specific. A captured example will mislead an operator on a different platform / state.
Wrappers, parsers, or scripts in any language that consume the tools' output. The output format is documented; users who want to script against it should read the live guide and write the parser against their installed version.
A samples/ or reference/ subtree. This is a thin loader for a documented DOCA artifact; substantive material lives on the public page and in the installed tool help.

Loading order

Read this skill first to confirm the user's question is in scope (the user wants to inspect or debug DPL programs on a BlueField DPU, not author, compile, or deploy DPL programs).
For what the tools inspect and change, the tool inventory, the read-only vs state-changing split, prerequisites, availability, support, and safety posture, see the capabilities companion.
For the documented invocations and the workflows — inspect, debug, modify, build, shell, plus the Command appendix — see the tasks companion.

doca-public-knowledge-map/SKILL.md — routing to the public DOCA Pipeline Language Developer Tools guide and the rest of the public DOCA documentation set. The canonical URL is reached via doca-public-knowledge-map ## DOCA tools (doca-public-knowledge-map/SKILL.md#doca-tools).
doca-version/SKILL.md — canonical DOCA version-handling rules. Version questions route here; the capabilities companion redirects to this skill for the body.
doca-setup/SKILL.md — env preparation, install verification, DPL Runtime Service container deployment, and the I have no install yet path with the public NGC DPL Development Container. This skill assumes its preconditions are satisfied.
doca-hardware-safety/SKILL.md — any step that changes hardware or firmware state routes here.
doca-structured-tools-contract/SKILL.md — the bundle's detect → prefer → fall back → report contract for structured helper tools. The Command appendix in the tasks companion honors this contract.
doca-debug/SKILL.md — the cross-cutting debug ladder. The DPL developer tools slot in at the runtime layer as inspection/debug tooling whose layered error surface escalates to the cross-cutting ladder when the cause is below DOCA.

Files

ATTRIBUTION.mdAdded by AutoGineer972 B
BENCHMARK.md10.7 KiB
CAPABILITIES.md10.3 KiB
LICENSELicence file added by AutoGineer11.2 KiB
SKILL.md14.7 KiB
SKILLCARD.yaml2.8 KiB
TASKS.md15.5 KiB
evals/evals.json3.5 KiB
skill-card.md4.9 KiB
skill.oms.sig5.2 KiB

BENCHMARK.md

Skill Benchmark: doca-pipeline-language-developer-tools

✅ Overall verdict: PASS — Recommended for publication

Publication Recommendation

Recommended for publication based on the completed evaluation evidence in this report.

Evaluation Metadata

Skill: doca-pipeline-language-developer-tools
Validation status: passed
Evaluation date: 2026-10-28
Evaluator version: 1.6.2
Agents: Claude Code (aws/anthropic/bedrock-claude-opus-5-5), Codex (openai/openai/gpt-6.1-sol)
Tasks: 5 evaluation tasks (3 positive, 2 negative)
Dataset digest: sha256:b8bbbcc052479714d9ea60279e534806882df2d3c9535b6b49a539ae925b6c5d (skill-evaluator-dataset-snapshot/1)
Attempts per task: 1
Environment: k8s-sandbox
Tier 2 evidence: required for publication
Tier 3 evidence: required for publication

Each task attempt ran in its own isolated sandbox pod.

Tier 1 LLM Provenance

SkillSpector provider, model, analyzer revision, and determinism evidence retained for this scan.

Skill	Provider	Adapter	Analyzer	Model	Revision	Determinism
doca-pipeline-language-developer-tools	nv_inference	openai	mcp_tool_poisoning	`aws/anthropic/bedrock-claude-opus-5-5`	`2.13.1`	nondeterministic / provider_defaults
doca-pipeline-language-developer-tools	nv_inference	openai	semantic_developer_intent	`aws/anthropic/bedrock-claude-opus-5-5`	`2.13.1`	nondeterministic / provider_defaults
doca-pipeline-language-developer-tools	nv_inference	openai	semantic_quality_policy	`aws/anthropic/bedrock-claude-opus-5-5`	`2.13.1`	nondeterministic / provider_defaults
doca-pipeline-language-developer-tools	nv_inference	openai	semantic_security_discovery	`aws/anthropic/bedrock-claude-opus-5-5`	`2.13.1`	nondeterministic / provider_defaults
doca-pipeline-language-developer-tools	nv_inference	openai	meta_analyzer	`aws/anthropic/bedrock-claude-opus-5-5`	`2.13.1`	nondeterministic / provider_defaults

Controls for doca-pipeline-language-developer-tools: temperature=not set (source=provider_default, forwarded=no); seed=not set (source=unset, forwarded=no); reasoning_effort=not set (source=provider_default, forwarded=no)

What This Report Answers

The three-tier evaluation checks whether the skill:

is safe to use;
produces correct answers;
is discovered and activated when needed;
helps the agent complete the user's goal and expected workflow; and
avoids wasted skill and tool usage.

Results at a Glance

Measure	Claude Code (Baseline → Skill Uplift)	Codex (Baseline → Skill Uplift)
Overall	92.2% — baseline ran, but no comparable score was available; uplift unavailable	92.5% — baseline ran, but no comparable score was available; uplift unavailable
Security	100.0% → 100.0% (±0.0 points)	90.0% → 100.0% (+10.0 points)
Correctness	40.0% → 88.0% (+48.0 points)	64.0% → 96.0% (+32.0 points)
Discoverability	100.0% — baseline ran, but no comparable score was available; uplift unavailable	93.3% — baseline ran, but no comparable score was available; uplift unavailable
Effectiveness	25.0% → 85.0% (+60.0 points)	35.0% → 100.0% (+65.0 points)
Efficiency	88.1% — baseline ran, but no comparable score was available; uplift unavailable	73.4% — baseline ran, but no comparable score was available; uplift unavailable

How to read this table: baseline is the same task attempted without the target skill. Scores are rounded to one decimal; threshold-adjacent values use additional precision so their displayed band matches the verdict. Uplift is derived from those displayed scores and shown in percentage points.

Example: 47.0% → 92.0% (+45.0 points) means the skill-assisted run scored 92.0%, 45.0 percentage points above its 47.0% no-skill baseline.

A partial dimension was calculated from only the available configured signals; review the detailed report before relying on it.

Comparison details

Shared-source comparisons use the same signals and weights on both conditions. They do not replace the full dimension scores above or change the publication verdict.

Agent	Dimension	Basis	Baseline → With skill (uplift)	Explanation
claude-code	Discoverability	Dimension comparison	N/A	Discoverability is not applicable to the no-skill baseline; there is no installed skill to discover.
claude-code	Efficiency	Token-only comparison	98.2% → 96.2% (-2.0 points)	Only Token Efficiency is comparable; the full dimension scores use different evaluator sources.
codex	Discoverability	Dimension comparison	N/A	Discoverability is not applicable to the no-skill baseline; there is no installed skill to discover.
codex	Efficiency	Token-only comparison	83.9% → 87.9% (+4.0 points)	Only Token Efficiency is comparable; the full dimension scores use different evaluator sources.

Token Usage

Actual Tier 3 execution usage is reported for every observed agent/case pair and both conditions.

Agent	Dataset case	With skill	Without skill	Delta	Change	Coverage
claude-code	All cases	722,551	539,273	+183,278	+33.99%	skill 5/5; base 5/5
claude-code	doca-pipeline-language-developer-tools.explicit-positive.v1	202,707	107,958	+94,749	+87.76%	skill 1/1; base 1/1
claude-code	doca-pipeline-language-developer-tools.implicit-positive.v1	202,732	108,088	+94,644	+87.56%	skill 1/1; base 1/1
claude-code	doca-pipeline-language-developer-tools.implicit-positive.v2	244,016	107,041	+136,975	+127.96%	skill 1/1; base 1/1
claude-code	doca-pipeline-language-developer-tools.negative.v1	36,780	108,108	-71,328	-65.98%	skill 1/1; base 1/1
claude-code	doca-pipeline-language-developer-tools.negative.v2	36,316	108,078	-71,762	-66.40%	skill 1/1; base 1/1
codex	All cases	329,109	249,047	+80,062	+32.15%	skill 5/5; base 5/5
codex	doca-pipeline-language-developer-tools.explicit-positive.v1	83,234	86,979	-3,745	-4.31%	skill 1/1; base 1/1
codex	doca-pipeline-language-developer-tools.implicit-positive.v1	85,304	30,711	+54,593	+177.76%	skill 1/1; base 1/1
codex	doca-pipeline-language-developer-tools.implicit-positive.v2	96,112	53,140	+42,972	+80.87%	skill 1/1; base 1/1
codex	doca-pipeline-language-developer-tools.negative.v1	34,448	48,010	-13,562	-28.25%	skill 1/1; base 1/1
codex	doca-pipeline-language-developer-tools.negative.v2	30,011	30,207	-196	-0.65%	skill 1/1; base 1/1
ALL AGENTS	Dataset aggregate	1,051,660	788,320	+263,340	+33.41%	skill 10/10; base 10/10

Prompt tokens include cached reads, so total tokens are prompt + completion (cached is not added twice). The Efficiency score uses (prompt - cached) + completion. Change is 100 * (with - without) / without; negative means fewer tokens. N/A comparisons include their reason in the coverage column; coverage is never estimated.

Tier Status

Tier	Purpose	Status	Evidence
Tier 1	Static validation	PASSED WITH OBSERVATIONS	11 validator(s); 19 finding(s)
Tier 2	Semantic deduplication	PASSED WITH OBSERVATIONS	2 validator(s); 3 finding(s)
Tier 3	Live agent evaluation	PASS	2 agent(s); 5 task(s)

Findings and Observations

<details> <summary>Show detailed findings and successful checks</summary>

CRITICAL CONTENT_DEDUP/llm_error: LLM analysis failed for a content cluster (skills/tools/doca-pipeline-language-developer-tools)
CRITICAL CONTENT_DEDUP/llm_error: LLM analysis failed for a content cluster (skills/tools/doca-pipeline-language-developer-tools)
HIGH DUPLICATE/duplicate: Duplicate content found within CAPABILITIES.md: "## Pattern overview" in CAPABILITIES.md (lines 33-49) vs "## Safety policy" in CAPABILITIES.md (lines 147-163) (CAPABILITIES.md:33)
MEDIUM QUALITY/quality_discoverability: Description uses first/second person (skills/tools/doca-pipeline-language-developer-tools/SKILL.md)
MEDIUM SCHEMA/metadata_key_style: Metadata key 'doca_version' is not kebab-case (skills/tools/doca-pipeline-language-developer-tools/SKILL.md)
17 additional finding(s) are available in the full evaluation artifacts.

</details>

Scoring Methodology

<details> <summary>Show dimension definitions, source signals, and thresholds</summary>

Dimension	Question	Scored signals
Security	Is it safe to use?	`security` (100%)
Correctness	Is the answer correct?	`accuracy` (100%)
Discoverability	Was the right skill loaded when needed?	`skill_execution` (100%)
Effectiveness	Did the skill help complete the task?	`goal_accuracy` (50%) + `behavior_check` (50%)
Efficiency	Did it avoid wasted tool calls and token usage?	`skill_efficiency` (50%) + `token_efficiency` (50%)

Dimension bands: PASS at 50% or above; NEUTRAL from 40% to below 50%; FAIL below 40%.
Overall Tier 3 lift: PASS at +5 points or more; FAIL at -10 points or less; values between those bands are NEUTRAL.
Overall verdict: PASS only when every configured dimension passes for at least one supported agent. Lift is reported as diagnostic evidence and does not override this gate.
The 50% attempt pass threshold is a separate per-task gate; it is not the dimension pass threshold.
Effectiveness is the equal-weight mean of goal completion (goal_accuracy) and expected workflow adherence (behavior_check).
Efficiency is 50% tool-call productivity (the backward-compatible skill_efficiency wire id) and 50% token_efficiency. Positive-case skill routing is scored under Discoverability, not Efficiency; a negative case without a routing target is N/A. N/A sources are omitted, remaining weights are renormalized, and the dimension is marked partial.

Signals present in this run:

security (Security): unsafe operations, secret leakage, and unauthorized access.
skill_execution (Skill Execution): whether the expected skill was selected, decoys were avoided, and the workflow executed.
skill_efficiency (Tool Productivity): tool-call productivity (legacy wire id; routing is scored under Discoverability).
accuracy (Accuracy): final-answer correctness against the reference answer.
goal_accuracy (Goal Accuracy): whether the user's goal was achieved.
behavior_check (Behavior Check): whether the expected workflow behavior was followed.
token_efficiency (Token Efficiency): actual uncached prompt plus completion usage (50% of Efficiency).

</details>

Freshness

Regenerate this benchmark when the skill, evaluation dataset, target agent/model, evaluator version, environment, or scoring policy changes.

Requirements and compatibility

Works with: Claude Code, Claude apps, OpenAI Codex, GitHub Copilot, Cursor, Gemini CLI, Kiro, OpenCode, Goose, Amp, Roo Code, JetBrains Junie, Factory Droid, Devin (ex-Windsurf), Mistral Vibe, OpenHands, Letta

“Works with” means the tool reads the open Agent Skills format; a skill written for one tool may use features others lack.

The author says: Requires a BlueField DPU configured to operate in DPU mode with the DPL Runtime Service container deployed and running, the dpl_rtd process active, and the DPL Development Container (publicly available on NGC) providing the developer tools. Debug sessions additionally require a DPL program compiled with the DPL Compiler debug flag -g enabled and loaded into dpl_rtd using the P4Runtime controller. The validation steps require a live DOCA install.

Security

What our scanner noticed· Medium risk

Large base64-encoded blobs

A static scan is not a guarantee. Skills are third-party content: review a skill before you let an agent use it.

Attribution and licence

Created by

NVIDIA

Publisher

NVIDIA (Official)

Source

NVIDIA/skills · skills/doca-pipeline-language-developer-tools · commit fce3cca

Licence

Apache License 2.0 (redistribution allowed)

LICENSE (added by AutoGineer from LICENSE-APACHE)
ATTRIBUTION.md (written by AutoGineer)

Synced

2026-11-09

AutoGineer is not affiliated with or endorsed by NVIDIA or NVIDIA.

Download ZIP

The ZIP holds the skill’s files, its licence files and an ATTRIBUTION.md. Download doca-pipeline-language-developer-tools.zip