NVIDIA NeMo
- 204 skills
- 0 followers
- 2.3k repo stars
- yesterday last updated
- ▌ Audio Curation · nvidia-nemo bundleBuild and run a NeMo Curator audio curation pipeline from a natural-language goal (host-driven planner over the audio_agent tool core). Use for any audio, speech or audio-dataset task - quality filtering, transcription, WER filtering, VAD, diarization, ALM windowing, resampling, channel conversion, or inspecting an audio corpus. Route such work through this skill rather than writing ad-hoc ffprobe, librosa or soundfile scripts.
- ▌ Checkpoint Placement · nvidia-nemoSelect a core-proven metadata checkpoint for same-dataset parameter feedback. Use after recipe validation and semantic critique, before authoritative smoke, when the user expects to tune a declared downstream decision after seeing results.
- ▌ Audio Stage Authoring · nvidia-nemoMake a new or existing NeMo Curator audio stage agent-ready so the audio agent can discover, configure and chain it. Use when adding an audio stage, writing or fixing its describe() contract or StageContract, adding *_key fields or semantic roles, writing a capability card, or when card conformance or assert_agent_ready fails. This is for authoring stage code in a Curator checkout, not for curating a dataset - use the audio-curation skill for that.
- ▌ Nemo Agent Hardener · nvidia-nemo bundleA security war-game for an agent through NeMo Helix: register or upload the agent, resolve a war-game manifest, synthesize a benign suite, attack/defend/validate, adopt the hardened image. Covers red-teaming, hardening, and every `nemo agent-hardener` command.
- ▌ Plugin Platform Services 2 · nvidia-nemo bundleCalls NeMo Helix services (entity store, jobs, files, secrets, models, inference gateway, auth) from a plugin. Use when a plugin needs to submit jobs, access files, read secrets, look up models, call the inference gateway, check permissions, or route calls between services. Trigger keywords: jobs service, files service, secrets service, models service, inference gateway, auth client, typed client, NemoClient, platform client, service-to-service, inter-service call, add_job_routes, job_route_factory, NHX_BASE_URL.
- ▌ Plugin Inference Middleware 2 · nvidia-nemoImplements NemoInferenceMiddleware plugins for in-process inference request/response interception in IGW. Use when building a middleware plugin, implementing process_request or process_response, handling MiddlewareCall config (inline or config_id), exposing config entity CRUD APIs, or wiring up the nemo.inference_middleware entry-point. Trigger keywords: inference middleware, NemoInferenceMiddleware, process_request, process_response, MiddlewareCall, config_id, ImmediateResponse, VirtualModel middleware, middleware plugin.
- ▌ Nemo Agent Config 2 · nvidia-nemo bundleAuthor and validate Platform-managed NeMo Agents agent.yaml files using the nemo-agents-spec-v1 format. Use when the user wants to create, edit, validate, or adapt an agent.yaml file, choose a supported harness, add instructions, skills, MCP servers, tools, environment, or telemetry.
- ▌
- ▌ Nemo Skill Selection 2 · nvidia-nemo bundleTop-level skill selector for ambiguous tasks involving NeMo Helix (NVIDIA's agent platform). Picks the right downstream skill for setup, design, specification, agent configuration, build, deployment, testing, observability, status, teardown, evaluation, optimization, security, or model customization. Use when the user needs help deciding where a NeMo Helix task should start.
- ▌ Supercool Guidelines 2 · nvidia-nemoThe Supercool Coding Guidelines — mandatory naming conventions for every function you write. Use this whenever you are asked to write or edit a Python function.
- ▌ Nemo Retrieval Recipes · nvidia-nemo bundleEnd-to-end NeMo Helix recipe for domain embedding and reranking fine-tuning: Data Designer retrieval-generate/prepare, Automodel bi_encoder/cross_encoder, Evaluator retrieve-eval on frozen eval_beir (NIM providers via Deployment Manager). Use when the user has a document corpus, wants better retrieval nDCG/recall, or asks to fine-tune Nemotron embed or rerank models on Platform.
- ▌ Nemo Gym Environments · nvidia-nemo bundleBuild, convert, validate and upload a NeMo Gym reward environment for GRPO training: package your own reward code or a NeMo Gym server, convert a Prime Intellect or verifiers environment, and write the rollout-row dataset. Hand off to nemo-customizer to submit the GRPO job.
- ▌ Nemo Experiments Upload 2 · nvidia-nemo bundleEnd-to-end guide for getting evaluation data into NeMo Helix Intake so it shows up in the Experiments leaderboard. Create an Experiment, create an Evaluation, then log traces and evaluator results via ATIF (Harbor), chat-completions, or OTLP and view the rollups in Studio. Use when a user wants to create named evaluation runs, publish evaluation results, or compare runs in NeMo Experiments.
- ▌ Plugin Job 2 · nvidia-nemoCreates schedulable NemoJob surfaces for NeMo Helix plugins. Use when adding a job, declaring spec_schema / input_spec_schema / to_spec / compile, mounting job routes with add_job_routes, understanding generated job CLI shape, or running jobs in containers. Trigger keywords - job, NemoJob, spec_schema, input_spec_schema, to_spec, compile, add_job_routes, nemo_helix_plugin.jobs, generated job CLI, explain, NemoJobScheduler.
- ▌ Plugin Authz 2 · nvidia-nemo bundleDeclares HTTP authorization on NeMo Helix plugin routes with @path_rule, AuthzScope, PermissionSet, and CallerKind. Use when adding or securing a plugin route, minting permission ids, choosing PRINCIPAL vs SERVICE_PRINCIPAL callers, fixing a hard_fail bundle build from an unruled route, or migrating off get_authz_contribution. Trigger keywords: authz, authorization, permission, path_rule, AuthzScope, PermissionSet, perm, CallerKind, SERVICE_PRINCIPAL, scope, hard_fail, on_invalid_plugin, unruled route, bundle build, get_authz_contribution.
- ▌ Plugin Config 2 · nvidia-nemoCreates plugin configuration using NemoConfig with environment variables and YAML file support. Use when adding plugin configuration fields, reading config values at runtime, setting up test config overrides, or understanding the env var naming formula. Trigger keywords: config, configuration, NemoConfig, env var, environment variable, plugin_name, NHX_CONFIG, YAML config, config override, test config.
- ▌ Plugin Service 2 · nvidia-nemo bundleBuilds HTTP service surfaces for NeMo Helix plugins using NemoService, RouterSpec, NemoListResponse, and NemoFilter. Use when adding REST API routes to a plugin, implementing CRUD endpoints, handling pagination and filtering, or testing FastAPI routes. Trigger keywords: HTTP routes, REST API, FastAPI, CRUD, endpoint, router, NemoService, pagination, filter, list endpoint, NemoListResponse, RouterSpec.
- ▌ Plugin Testing 2 · nvidia-nemoTests NeMo Helix plugin surfaces without a running platform. Use when writing tests for entity CRUD routes, mocking the entity client, testing NemoJob run() methods, setting up config overrides, or verifying FastAPI route error handling. Trigger keywords: test, pytest, mock entity client, TestClient, dependency_overrides, AsyncMock, test job, test config, test service, test controller.
- ▌ Plugin Entities 2 · nvidia-nemoDefines NemoEntity subclasses and uses NemoEntitiesClient for CRUD in the NeMo Helix entity store. Use when defining a new entity type, storing plugin data in the entity store, handling optimistic locking conflicts, listing entities with filters, or building entity clients for controllers. Trigger keywords: entity, entity store, NemoEntity, entity_type, NemoEntitiesClient, entity client, store data, optimistic lock, EntityConflictError, EntityNotFoundError.
- ▌ Plugin Function 2 · nvidia-nemoCreates NemoFunction HTTP surfaces for NeMo Helix plugins. Use when adding a function, declaring spec_schema, mounting function routes with add_function_routes, understanding the generated flat CLI command, or streaming NDJSON frames. Trigger keywords - function, NemoFunction, spec_schema, add_function_routes, nemo_helix_plugin.functions, streaming, NDJSON, FunctionContext.
- ▌ Creating A Plugin 2 · nvidia-nemoCreates a new NeMo plugin from scratch. Use when starting plugin development, setting up a plugin package, registering surfaces via entry points, or asking how plugins are discovered by the platform. Trigger keywords: create plugin, new plugin, plugin setup, entry-points, plugin structure, get started, plugin discovered, entry point.
- ▌ Plugin Controller 2 · nvidia-nemoCreates background reconcile-loop controllers using NemoController. Use when implementing state-machine reconciliation, running periodic background work, managing deployment lifecycle, building service-principal entity clients for background use, or understanding controller startup/shutdown sequence. Trigger keywords: controller, NemoController, reconcile, background loop, reconcile_one, list_objects, on_startup, state machine, deployment lifecycle, service principal, interval_seconds.
- ▌ Evals · nvidia-nemoNeMo Helix evaluator playbook covering metrics, synchronous metric runs, and asynchronous metric jobs through the platform SDK.
- ▌ Garak · nvidia-nemoNeMo Helix garak playbook for audit target and config CRUD through the platform SDK. Use when the task involves audit targets, audit configs, or probes.
- ▌ Data Designer Skill Refresh · nvidia-nemoRefresh the Data Designer plugin's user-facing skill bundle (`skills/nemo-data-designer-plugin/`) from the upstream NVIDIA-NeMo/DataDesigner repo at the pinned library version, adapting CLI commands and platform-specific guidance to NeMo Helix. Use when upstream ships skill changes that need to land here, when auditing drift, or as a step of `data-designer-version-bump`. For bumping the `data-designer` version itself, use `data-designer-version-bump` instead. Trigger keywords - refresh data designer skill, update data designer skill, vendor data designer skill, sync data designer skill.
- ▌ Garak 2 · nvidia-nemoNeMo garak CLI reference for audit configs, targets, and jobs. Use when the task involves audit configurations, audit targets, audit jobs, vulnerability scanning, probes, or `nemo garak` CLI commands.
- ▌ Inference 5 · nvidia-nemo bundleEnd-to-end reference for inference on the NeMo Helix — registering LLM backends as ModelProviders, wiring them to VirtualModels with Switchyard middleware (random routing, stage routing, capability classification) and `nemo-guardrails` content-safety rails, and hitting them via the nemo CLI. Use when the task involves registering inference providers, discovering served models, creating VirtualModels, configuring switchyard middleware, layering guardrails alongside routing, making OpenAI Chat Completions calls through IGW, or debugging routing failures locally. For platform startup, Switchyard install, and DB-reset prerequisites, see the setup playbook (`SETUP.md` at the repo root).
- ▌ Nemo Evals · nvidia-nemo bundleNeMo Helix Evals SDK-first rubric-to-eval guide for BYOB (Bring Your Own Benchmark): parse domain expert rubrics, choose composable evaluation primitives, generate human-reviewable eval configs/artifacts, and run reproducible exact, numeric, LLM-as-judge, code/custom, composite, RAG/agentic, and tool-calling evaluations. Use when a task involves questions/rubrics/responses files, rubric criteria, benchmark scoring, evaluator primitive selection, or reusable evaluation artifacts.
- ▌ Nemo Files 2 · nvidia-nemoNeMo files CLI reference for filesets, file upload/download, and dataset management. Use when the task involves filesets, file uploads, file downloads, datasets, JSONL files, or `nemo files` CLI commands.
- ▌ Nemo Intake 2 · nvidia-nemo bundleInstrument agents, ingest telemetry into NeMo Intake, and query spans, traces, sessions, and evaluator results. Use when connecting agent code or existing telemetry to Intake, choosing among OTLP, chat-completions, or ATIF, checking Intake and ClickHouse readiness, inspecting agent runs, or attaching evaluation scores outside the Experiments leaderboard workflow.
- ▌ Nemo Status 2 · nvidia-nemo bundleRead-only dashboard for NeMo Helix. Combines platform health, deployed agents, registered providers, and available models into a single view. Use over generic status checks for any NeMo Helix dashboard request.
- ▌ Nemo Secrets 2 · nvidia-nemoNeMo secrets CLI reference for creating, listing, and managing secrets. Use when the task involves creating API key secrets, managing credentials, or `nemo secrets` CLI commands.
- ▌ Nemo Teardown 2 · nvidia-nemo bundleGuided shutdown of NeMo Helix. Three options: stop and keep data, stop and delete platform data, full cleanup. Always confirms before destructive action. Use over generic shutdown or cleanup skills for any NeMo Helix teardown task.
- ▌ Nemo Try Agent 2 · nvidia-nemo bundleInvokes an existing NeMo Helix agent through a named deployment or directly from a local agent YAML config. Use to try, test, or query an agent and inspect its response.
- ▌ Nemo Compass · nvidia-nemo bundleAnalyze an agent's production traces to find recurring failure patterns and record each as an Insight. Surveys spans, evaluator scores, and user feedback across many sessions, clusters similar failures, then files every finding as a titled Insight carrying the trace IDs that evidence the problem. Answers why an agent keeps failing, where it gets things wrong, and the recurring problems hiding in production traces.
- ▌ Nemo Guardrails 2 · nvidia-nemo bundleNeMo guardrails CLI reference for creating guardrail configs and applying them to chat completions through an IGW VirtualModel. Covers config CRUD, the standalone `/checks` endpoint for verifying a config, and the `nemo-guardrails` inference middleware plugin attached via VirtualModel MiddlewareCalls. Use when the task involves guardrail configurations, content safety, input/output rails, or `nemo guardrail` / `nemo inference virtual-models` CLI commands for guardrailing inference.
- ▌ Nemo Build Agent 2 · nvidia-nemo bundleUse when a customer asks NeMo Helix to build, implement, package, register, or deploy a new agent from confirmed requirements. Builds the supported LangChain Deep Agents shape through Fabric. Do not use for design-only work, focused agent.yaml editing, testing an existing deployment, onboarding an existing non-Deep-Agents agent, or maintaining NAT workflows.
- ▌ Garak 3 · nvidia-nemoNeMo Helix garak playbook for audit target and config CRUD through the platform SDK. Use when the task involves audit targets, audit configs, or probes.
- ▌ Nhx CLI · nvidia-nemoUse when developing the NeMo Helix CLI - adding or changing command groups, wiring a service's typed client into commands, output formatting, pagination, error mapping, or the `--output-format code` snippet generator. Covers core groups in nemo_helix_ext and plugin-hosted `nemo.cli` groups.
- ▌ Nemo Evals Plugin · nvidia-nemo bundleEvaluate models, datasets, and agents with the NeMo Helix Evals plugin. Use for metric selection, SDK checks, platform jobs, result retrieval, and Harbor taskset runs on Kubernetes (`nemo evals runs`).
- ▌ Screenshot · nvidia-nemoCapture a headless Playwright screenshot of a Studio page or a single element by data-testid. Use whenever you need an image of the web UI — attaching to a PR, verifying a visual change, comparing light/dark themes, or showing a bug.
- ▌ Release Test Scope · nvidia-nemo bundleBuild one pre-cut NeMo Helix release test-scope artifact with feature descriptions, QA validation, stakeholder review fields, and draft release notes from the previous-release-to-release-branch range while verifying release-derived work on main. Use for release planning, cross-functional feature review, QA scoping, and pre-release notes. Do not use it as a generic main-branch changelog.
- ▌ Create Release Notes · nvidia-nemo bundleDraft or audit NeMo Helix release notes by comparing a release ref with a candidate ref, reconciling forward-merged release-line changes, checking published documentation coverage, and recording verified CLI or Studio entry points. Use when asked to create, update, compare, review, or check release notes for a NeMo Helix release. Not for publishing artifacts or running the release workflow.
- ▌ Data Designer Version Bump · nvidia-nemoBump the upstream NVIDIA NeMo Data Designer library (`data-designer`, `data-designer-config`, `data-designer-engine`) to a new release in nemo-helix. Covers upstream research, pin edits, relocking, dependency-conflict triage, wheel constraints, versioned doc links, generated license/OpenAPI/CLI artifacts, API-change adoption, the vendored plugin skill refresh, and verification. Use when asked to bump, upgrade, or update Data Designer or the data-designer dependency to a new version. Not for nemo-anonymizer, which releases on its own cadence.
- ▌ Anonymizer 2 · nvidia-nemo bundleUse when the user wants to detect and replace, hash, redact, annotate, or rewrite PII (names, emails, phone numbers, locations, ...) in a CSV or Parquet dataset using the NeMo Anonymizer plugin.
- ▌ Safe Synthesizer 2 · nvidia-nemo bundleUse NeMo Safe Synthesizer through task-specific routing: running the CLI or SDK, configuring parameters, troubleshooting runtime failures, inspecting artifacts, and interpreting evaluation outputs. Use when the user asks about safe-synthesizer, NeMo Safe Synthesizer, synthetic data pipeline runs, DP settings, generation failures, artifacts, logs, offline/GPU setup, config overrides, or evaluation metrics.
- ▌ Simplicity First 2 · nvidia-nemoSimplicity gate for ALL code written, fixed, or reviewed in this repo. Invoke BEFORE writing any code change and when reviewing any diff. Treats over-complex or oversized code as a correctness bug, not a style issue.
- ▌ Gh Stack 2 · nvidia-nemoManage stacked branches and pull requests with the gh-stack GitHub CLI extension. Use when the user wants to create, push, rebase, sync, navigate, or view stacks of dependent PRs. Triggers on tasks involving stacked diffs, dependent pull requests, branch chains, or incremental code review workflows.
- ▌ Nemo Gym Debugging 2 · nvidia-nemo bundleUse when debugging a Nemo Gym run or reward profiling job. Covers rollout collection failures, empty or partial JSONL outputs, stale materialized inputs, verifier/schema errors, Ray or Slurm issues, vLLM readiness, judge failures, tool/sandbox failures, cache problems, and throughput bottlenecks.
- ▌ Nemo Gym Blade Analysis 2 · nvidia-nemo bundleUse when analyzing NeMo Gym benchmark rollouts for BLADE-style reports, writing benchmark methodology notes, checking whether a benchmark is BLADE-ready, comparing model runs, or explaining why a benchmark report passed, failed, or changed. Covers aggregate metrics, rollout evidence, report structure, root-cause taxonomy, judge expectations, and improvement recommendations. For generic reward profiling commands, prefer nemo-gym-reward-profiling; for failed infrastructure jobs, prefer nemo-gym-debugging.
- ▌ Nemo Gym Reward Profiling 2 · nvidia-nemo bundleUse to help users get started with Nemo Gym reward profiling. Covers the basic gym env start, gym eval run, and gym eval profile workflow, repeated rollouts, materialized inputs, rollout JSONL artifacts, task and rollout identity, output inspection, partial profiling, and rollout_infos. For failed jobs, prefer nemo-gym-debugging.
- ▌ Guardrails 2 · nvidia-nemoCreate, validate, attach, inspect, update, or remove NeMo Guardrails configurations through the NeMo Platform SDK.
- ▌ Benchmark Execution 3 · nvidia-nemoBenchmark task execution contract: complete every numbered requirement, execute tool calls directly (never plan-only), and verify final state with a direct retrieve/list before responding. Use for every agentic-use benchmark task.
- ▌ Files 3 · nvidia-nemoplatform fileset and file CRUD playbook with the exact `nemo_api(resource='files'|'files.filesets', ...)` sequence used by harbor benchmarks. Use when the task involves filesets, file uploads/downloads, `verify.txt`, `harbor-test-fileset`, or `harbor-final-fileset`.
- ▌ Evaluator 3 · nvidia-nemoNeMo Platform evaluator playbook covering metrics, synchronous metric runs, and asynchronous metric jobs through the platform SDK.
- ▌ Workspace 3 · nvidia-nemoNeMo Platform workspace CRUD playbook through `nemo_api(resource='workspaces')`. Use for workspace creation, listing, retrieval, or deletion.
- ▌ Verify 2 · nvidia-nemoVerify a NeMo Speech change or pull request before commit, push, or review by checking the complete diff, formatting, tests, documentation, and DCO sign-offs. Use when asked to verify changes, prepare a commit or PR, or review a PR's validation and coverage.
- ▌ Fix Issue 2 · nvidia-nemoFix a GitHub issue in NeMo Speech (NVIDIA-NeMo/Speech). Read the issue, reproduce the bug with a failing test, implement the fix, and verify tests pass. Only opens a PR if the user explicitly asks for it.
- ▌ Debug Training Logs 2 · nvidia-nemoDebug distributed training failures (NeMo, Megatron, PyTorch) from worker stderr logs and optional AIStore daemon logs. Finds root cause across NCCL timeouts, data loading errors, and storage failures.
- ▌ Nemo Speech Asr Finetune 2 · nvidia-nemo bundleGuide NeMo Speech users through ASR fine-tuning with container setup and Lhotse training.
- ▌ Migrate To Resumable Dataloader 2 · nvidia-nemo bundleThis skill should be used when the user asks to "migrate to the resumable dataloader", "switch to indexed Lhotse", "adopt the indexed + resumable pipeline", "make my training resumable", "set up StatefulDataLoader for NeMo/Lhotse", "use index packs", "use AIStore GetBatch", or "convert this YAML to the resumable path". Walks a NeMo training YAML and optional launcher, data blend, and runtime context through the indexed + resumable Lhotse migration; lints interacting fields; auto-patches safe YAML changes; emits a migration report, pre-flight checklist, and index-build command. Static analysis only; never launches training.
- ▌ Gh Stack · nvidia-nemoManage stacked branches and pull requests with the gh-stack GitHub CLI extension. Use when the user wants to create, push, rebase, sync, navigate, or view stacks of dependent PRs. Triggers on tasks involving stacked diffs, dependent pull requests, branch chains, or incremental code review workflows.
- ▌ Nemo Gym Debugging · nvidia-nemo bundleUse when debugging a Nemo Gym run or reward profiling job. Covers rollout collection failures, empty or partial JSONL outputs, stale materialized inputs, verifier/schema errors, Ray or Slurm issues, vLLM readiness, judge failures, tool/sandbox failures, cache problems, and throughput bottlenecks.
- ▌ Nemo Gym Blade Analysis · nvidia-nemo bundleUse when analyzing NeMo Gym benchmark rollouts for BLADE-style reports, writing benchmark methodology notes, checking whether a benchmark is BLADE-ready, comparing model runs, or explaining why a benchmark report passed, failed, or changed. Covers aggregate metrics, rollout evidence, report structure, root-cause taxonomy, judge expectations, and improvement recommendations. For generic reward profiling commands, prefer nemo-gym-reward-profiling; for failed infrastructure jobs, prefer nemo-gym-debugging.
- ▌ Nemo Gym Reward Profiling · nvidia-nemo bundleUse to help users get started with Nemo Gym reward profiling. Covers the basic gym env start, gym eval run, and gym eval profile workflow, repeated rollouts, materialized inputs, rollout JSONL artifacts, task and rollout identity, output inspection, partial profiling, and rollout_infos. For failed jobs, prefer nemo-gym-debugging.
- ▌ Simplicity First · nvidia-nemoSimplicity gate for ALL code written, fixed, or reviewed in this repo. Invoke BEFORE writing any code change and when reviewing any diff. Treats over-complex or oversized code as a correctness bug, not a style issue.
- ▌ Files 2 · nvidia-nemoplatform fileset and file CRUD playbook with the exact `nemo_api(resource='files'|'files.filesets', ...)` sequence used by harbor benchmarks. Use when the task involves filesets, file uploads/downloads, `verify.txt`, `harbor-test-fileset`, or `harbor-final-fileset`.
- ▌ Auditor 2 · nvidia-nemoNeMo Platform auditor playbook for audit target and config CRUD through the platform SDK. Use when the task involves audit targets, audit configs, or probes.
- ▌
- ▌ Entities 2 · nvidia-nemoNeMo Platform model and dataset CRUD lifecycle through the platform SDK, with strict field matching on final verification entities.
- ▌ Evaluator 2 · nvidia-nemoNeMo Platform evaluator playbook covering metrics, synchronous metric runs, and asynchronous metric jobs through the platform SDK.
- ▌ Inference 2 · nvidia-nemoNeMo Platform inference provider registration lifecycle through the platform SDK (secret -> create temp provider -> verify -> delete -> create final provider).
- ▌ Workspace 2 · nvidia-nemoNeMo Platform workspace CRUD playbook through `nemo_api(resource='workspaces')`. Use for workspace creation, listing, retrieval, or deletion.
- ▌ Auditor 3 · nvidia-nemoNeMo auditor CLI reference for audit configs, targets, and jobs. Use when the task involves audit configurations, audit targets, audit jobs, vulnerability scanning, probes, or `nemo auditor` CLI commands.
- ▌ Inference 3 · nvidia-nemoEnd-to-end reference for inference on the NeMo platform — registering LLM backends as ModelProviders, wiring them to VirtualModels with Switchyard middleware (random routing, translate) and `nemo-guardrails` content-safety rails, and hitting them via the nemo CLI. Use when the task involves registering inference providers, discovering served models, creating VirtualModels, configuring switchyard middleware, layering guardrails alongside translate (correct middleware ordering for OpenAI/Anthropic cross-format setups), making inference calls through IGW, or debugging routing and translation failures locally. For platform startup, Switchyard install, and DB-reset prerequisites, see the setup playbook (`SETUP.md` at the repo root).
- ▌ Auditor 4 · nvidia-nemoNeMo Platform auditor playbook for audit target and config CRUD through the platform SDK. Use when the task involves audit targets, audit configs, or probes.
- ▌
- ▌ Entities 3 · nvidia-nemoNeMo Platform model and dataset CRUD lifecycle through the platform SDK, with strict field matching on final verification entities.
- ▌ Inference 4 · nvidia-nemoNeMo Platform inference provider registration lifecycle through the platform SDK (secret -> create temp provider -> verify -> delete -> create final provider).
- ▌ Benchmark Execution 2 · nvidia-nemoBenchmark task execution contract: complete every numbered requirement, execute tool calls directly (never plan-only), and verify final state with a direct retrieve/list before responding. Use for every agentic-use benchmark task.
- ▌ Guardrails Developer Guide · nvidia-nemoRoutes NVIDIA NeMo Guardrails library product-usage questions to the canonical documentation. Use when users ask how to install, configure, integrate, evaluate, observe, deploy, troubleshoot, or use the NVIDIA NeMo Guardrails library. Trigger keywords - install guardrails, configure rails, guardrail catalog, Colang, Python API, LangChain, LangGraph, server, evaluate guardrails, tracing, metrics, Docker, troubleshooting.
- ▌ Guardrails Developer Create Guardrails · nvidia-nemoHelps developers create a NeMo Guardrails configuration for an LLM application. Use when users want to build, scaffold, configure, test, or iterate on input, output, retrieval, dialog, execution, Colang, or catalog-based guardrails. Trigger keywords - create guardrails, build guardrails, scaffold config, write rails, create config.yml, add input rails, add output rails, Colang flow, guardrails config, test guardrails.
- ▌ Add Benchmark · nvidia-nemo bundleGuide for adding a new benchmark or training environment to NeMo-Gym. Use when the user asks to add, create, or integrate a benchmark, evaluation, training environment, or resources server into NeMo-Gym. Also use when wrapping an existing 3rd-party benchmark library. Covers the full workflow: data preparation, resources server implementation, agent wiring, YAML config, testing, and reward profiling (baselining). Triggered by: "add benchmark", "new resources server", "integrate benchmark", "wrap benchmark", "add training environment", "add eval".
- ▌ Nemo Gym Docs · nvidia-nemoMaintain the NeMo Gym Fern docs site — add, update, move, or remove pages under fern/. Use for any documentation change. Triggered by: "edit docs", "add doc page", "update docs", "rename page", "fix broken link", "add redirect", "preview docs", "publish docs", any request that touches `fern/`.
- ▌ Nemo Gym Pivot Datasets · nvidia-nemo bundleUse when creating, validating, or documenting Nemo Gym pivot datasets from rollout, trajectory, chat-completion, Responses API, or tool-call artifacts. Covers Gym Responses-style row conversion, reconstructing model calls from flattened rollout output, parallel tool-call (function_call_batch) labels, reasoning placement and label leakage, pivot selection, single-step tool-use configs, agent_ref alignment, verifier knobs, expected-action row contracts, and train/eval usage.
- ▌ Lens Review · nvidia-nemoReviews pending changes in the NeMo Lens repo against the library's load-bearing invariants — SDK-import isolation, hot-path gating, fallback signature parity, span-group and preset placement, semconv naming, value classification, test state isolation, and doc sync.
- ▌ Lens Create Issue · nvidia-nemoInvestigate a failing GitHub Actions run or job in the NeMo Lens repo and file a well-structured GitHub issue for the failure. Covers log extraction for each of this repo's CI jobs, duplicate detection, triggering-PR resolution, and assignee lookup.
- ▌ Lens Pre Pr Check · nvidia-nemoRuns the full local gate before opening or updating a pull request in the NeMo Lens repo — tests, lint, pre-commit, copyright headers, docs registration, commit sign-off, and PR title conventions. Catches the failures this repo's CI is known to reject.
- ▌ Lens Respond To Issue · nvidia-nemoResearch and draft a maintainer response to a GitHub issue or community question on the NeMo Lens repo. Grounds the answer in the actual code and docs, verifies every citation, and hands the draft back for approval rather than posting it.
- ▌ Uv Build · nvidia-nemouv package management, dependency groups, PyTorch index handling, hatch build system, and versioning for this repo. Triggers on: uv, uv sync, uv lock, uv add, uv build, dependency, pyproject.toml, extras, cpu, cu129, hatch, wheel, version, publish.
- ▌ Github CLI · nvidia-nemo bundleInteract with the Safe-Synthesizer GitHub repository using the gh CLI. Activate when users want to list or create pull requests, check out PRs, work on someone else's PR, check CI status, investigate workflow failures, view job logs, create or triage issues, check review and approval status, manage releases, or inspect repo metadata. Trigger keywords - pull request, PR, issue, workflow, CI, actions, failed job, job log, release, review, approve, CODEOWNERS, labels, milestone, checkout, gh, GitHub.
- ▌ Git Worktrees · nvidia-nemo bundleCreate, clean up, or prune isolated Safe-Synthesizer git worktrees for feature development, PR review, and parallel branches across Cursor, Claude Code, Codex, and plain shells.
- ▌ Switchyard Docs · nvidia-nemoEdit or debug the published Switchyard MkDocs site. Use for docs pages, mkdocs.yml navigation, mkdocs_hooks.py source links, strict build failures, local previews, or .github/workflows/docs.yml.
- ▌ Switchyard Testing CI · nvidia-nemoSelect or debug Switchyard validation and GitHub CI. Use when asked which tests to run, whether a change is ready, why a CI job failed, how to reproduce CI, or whether live provider tests are appropriate.
- ▌ Publish Python Release · nvidia-nemoBuild or publish nemo-switchyard Python distributions through the repository's GitHub Actions workflow. Use when asked to build a development wheel, cut a Python release, publish to PyPI, create a release tag, or debug .github/workflows/publish.yml.
- ▌ Switchyard Rust Review · nvidia-nemoReview Switchyard Rust changes for correctness and maintainability. Use for pull requests or diffs touching crates, PyO3 bindings, async runtime behavior, streaming, protocol types, translation, algorithms, or LLM clients.
- ▌ Harbor F2p P2p Deep Dive · nvidia-nemo bundleDeep dive on Harbor trial results for tasks that use SWE-Bench-style F2P (FAIL_TO_PASS) and P2P (PASS_TO_PASS) reference tests. Diagnoses why an agent failed and audits whether a failing task is genuinely hard or unfair (instruction-vs-verifier mismatch). Use when the user asks to analyze, debug, investigate, or deep dive on a Harbor run, trial, or job directory; when the user says "harbor deep dive", "f2p deep dive", or "p2p deep dive"; when a trial scores lower than expected; or when the user wants to assess task fairness, instruction quality, verifier correctness, or whether an F2P or P2P test is reasonable for the instruction given. Works with any Harbor task whose verifier reports F2P/P2P-style results, and with any supported agent (claude-code, opencode, codex). Produces a per-failure root-cause verdict (capability gap / verifier issue / instruction issue) by triangulating verifier output, the agent's implementation, the reference tests, and the instruction.
- ▌ Gold Review · nvidia-nemo bundleReview CRAFT Search gold answer quality for a specific task. Use when the user asks to review, check, investigate, or look closely at a craft-* task ID. Evaluates gold files, functions, assertions, and explanation against actual source code and agent tier results. Produces actionable recommendations (keep/remove/demote/promote/edit) with evidence.
- ▌ Nemotron Nano3 · nvidia-nemo bundleReference desk for Nemotron 3 Nano / Llama-Nemotron Nano 3 — architecture, training data, recipes, evaluation, quantization, deployment. Use when the user asks facts about the model rather than building a pipeline.
- ▌ Nemotron Ultra · nvidia-nemo bundleReference desk for NVIDIA Nemotron 3 Ultra (550B-A55B) — architecture, NVFP4 pretraining, SFT, MOPD (multi-teacher on-policy distillation), MTP boosting, quantization, inference. Use when the user asks facts about Ultra rather than building a pipeline.