Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx agentmods add skills/nvidia-nemo/megatron-bridge/adding-model-supportnpx skills add NVIDIA-NeMo/Megatron-Bridge --skill adding-model-supportgit clone --depth 1 https://github.com/NVIDIA-NeMo/Megatron-BridgeWhat it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5 | $0.00036 | $0.04973 |
| Opus 5 | $0.00018 | $0.02486 |
| Sonnet 5 | $0.00007 | $0.00995 |
| Haiku 4.5 | $0.00004 | $0.00497 |
Grade A, and why
adding-model-support scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 3d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 463 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Adding New Model Support in Megatron-Bridge
Phase 1: Discovery
Step 1 — Get the HF model link
Ask the user for the HuggingFace model link (e.g. https://huggingface.co/Qwen/Qwen3.5-VL-27B).
If the model is not public, ask the user to provide the config.json file directly.
Step 2 — Fetch and analyze config.json
Read the model's config.json from HuggingFace (or from the user-provided file). Key fields to extract:
model_type— used for@register_bridge(model_type=...)architectures— the HF model class name (used forsource=...in registration)tie_word_embeddings— critical for weight tying- Architecture fields:
num_hidden_layers,hidden_size,intermediate_size,num_attention_heads,num_key_value_heads,vocab_size,max_position_embeddings,rope_theta, etc. - MoE fields (if present):
num_local_experts,num_experts_per_tok,moe_intermediate_size - MLA fields (if present):
q_lora_rank,kv_lora_rank,qk_nope_head_dim,qk_rope_head_dim
If there are config fields you don't recognize from previously supported models (check CONFIG_MAPPING in model_bridge.py and existing bridges), this likely indicates a new architectural block (e.g., a novel attention variant, custom normalization, or a new layer type). Ask the user to provide the HuggingFace modeling_*.py implementation of that block so you can understand the computation and create the correct Megatron-side mapping or custom module.
Step 3 — Determine VLM vs LLM
VLM (Vision-Language Model) if config.json contains:
text_configANDvision_configsub-configs- Note: VLMs may or may not have "VL" in the name
LLM (Text-only) if:
- No
text_config/vision_config - Single flat config for the language model
This distinction affects:
- Which files to create (VLMs need a model.py combining vision + language)
- Where to read config fields from (
text_configvs top-level for VLMs) - Test patterns (VLMs need vision inputs in functional tests)
What ships with it
4 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 3d ago First seen · 463 lines · 36 tokens per session scan A ad38996e271d
adding-model-support is a skill published in the GitHub repository NVIDIA-NeMo/Megatron-Bridge (893 stars, last pushed yesterday), licensed Apache-2.0. It adds 36 tokens to every session and 4,973 once invoked, about $0.0002 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other skills, from other repositories
agent-platform-model-registry
Agent Platform Model Registry Management. Use when you need to upload, list, describe, update, or delete machine learning models (and their versions) in the Agent Platform Model Registry. Don't use for model training, model deployment to endpoints, or managing non-Agent Platform models.
bigquery-ai-ml
Leverages BigQuery's built-in machine learning and GenAI capabilities for advanced data analytics. Use when you need to write SQL queries that perform time-series forecasting, predict values, detect outliers or anomalies, find key drivers, perform semantic search or vector search, classify text, calculate similarity…
ondb
A logical analysis and reasoning tool for AI. Use when decomposing documents into structured knowledge, querying entities and relations, validating consistency, or indexing files. Trigger on "remember", "what do I know about", "link X to Y", "show dependencies", "analyze this document", entity CRUD, or cross-skill…
neuron-evaluation-engineer
Create and run AI evaluations with datasets, assertions, and output drivers in Neuron AI. Use this skill whenever the user mentions evaluation, testing AI systems, creating evaluators, dataset-driven testing, assertion-based validation, or wants to measure AI system performance. Also trigger for tasks involving…
mixed-precision
Use FP16/BF16 mixed precision to accelerate training and reduce memory. Use when optimizing GPU performance.
sparse-autoencoder-training
Provides guidance for training and analyzing Sparse Autoencoders (SAEs) using SAELens to decompose neural network activations into interpretable features. Use when discovering interpretable features, analyzing superposition, or studying monosemantic representations in language models.