Getting it into your agent
It runs from inside its repository, so the clone comes first — what it calls does not travel with the file alone.
git clone --depth 1 https://github.com/NVIDIA-NeMo/Megatron-Bridgenpx agentmods add skills/nvidia-nemo/megatron-bridge/create-model-verification-cardWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/nvidia-nemo/megatron-bridge/create-model-verification-card)<a href="https://agentmods.dev/skills/nvidia-nemo/megatron-bridge/create-model-verification-card"><img src="https://agentmods.dev/badge/skills/nvidia-nemo/megatron-bridge/create-model-verification-card/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/nvidia-nemo/megatron-bridge/create-model-verification-card"><img src="https://agentmods.dev/badge/skills/nvidia-nemo/megatron-bridge/create-model-verification-card.svg" alt="Reviewed on agentmods" width="80" height="20"></a>- NVIDIA SkillSpector pass
What 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 5.1 | $0.00120 | $0.12102 |
| Opus 5.5 | $0.00048 | $0.04841 |
| Sonnet 5.5 | $0.00024 | $0.02420 |
| Haiku 4.5 | $0.00012 | $0.01210 |
Grade A, and why
create-model-verification-card 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 11d 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 — 989 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Create Model Verification Card
Create examples/model_verification_cards/<model-slug>/card.yaml from public
model facts and verified commands. Keep the card small enough for an agent to
scan without interpreting logs or reconstructing the execution environment.
Use the repository resources
Treat verification scripts, validators, and launchers as shared infrastructure. Do not modify them merely to make one model card pass. Any such change requires a clear, documented, reusable reason: identify the existing behavior that is insufficient, the affected public workflows or models, why an existing maintained path cannot be used, and add focused backward-compatible tests. Record the justification in the PR description or commit. If the need is model-specific or the reason is not clear, leave the affected verification item unverified instead of adding a card-only workaround.
- Validate the result with scripts/validate_card.py.
- Verify deterministic HF output with scripts/verify_hf_inference.py.
- Use the inventory and field rules below as the format contract. Do not infer model-specific settings from another family or variant.
Do not add a README, evidence blobs, log excerpts, runtime setup, or scheduler metadata to the skill or card.
Workflow
1. Pull facts before drafting
Read the model implementation, public HF config, conversion bridge, recipes, tests, examples, and the exact revision being verified. Determine which modes exist; do not infer support from a family name or another model size.
Record the public HF model name in commands, its immutable revision in
model.hf_revision, and the minimum supported Transformers version in
model.min_transformers_version. Do not introduce an HF snapshot path.
Record only two execution-environment facts: the public base container
identifier and the exact Bridge commit used for verification. Put them in
verification_environment.base_container and
verification_environment.bridge_commit. If the run mounted a checkout over
the container, record the mounted checkout commit. Never substitute a private
image path for the public base container identifier.
What ships with it
2 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.
- 11d ago Changed · +31 lines 7f41d749a449
- 13d ago Changed · +10 lines 0e680291d330
- 23d ago Changed · +23 lines 0babdb319dc1
- 27d ago Changed · +7 lines 663879a46012
- 1mo ago Changed · +33 lines 5c9ba2533cc3
- 1mo ago First seen · 885 lines · 120 tokens per session scan A 036f9dee4745
create-model-verification-card is a skill published in the GitHub repository NVIDIA-NeMo/Megatron-Bridge (970 stars, last pushed today), licensed Apache-2.0. It adds 120 tokens to every session and 12,102 once invoked, about $0.0005 per session on Opus 5.5. 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
research-engineer
An uncompromising Academic Research Engineer. Operates with absolute scientific rigor, objective criticism, and zero flair. Focuses on theoretical correctness, formal verification, and optimal implementation across any required technology.
tika-eval-compare
Compare extracts from two Tika builds over a corpus to detect regressions in content, encoding, exceptions, and embedded-document handling. Use for "compare before/after extracts", "eval this change against the corpus".
cua-driver
Use Cua Driver for desktop or browser tasks that are awkward or unavailable through Bash/APIs, or when the user explicitly wants GUI interaction: app testing, visual bug reproduction, form filling, calendar entry, screenshots, and demo recording. Also covers Cua setup; not OpenAI Codex Computer Use or web research.
jetson-validate-image
Use after jetson-flash-image to run static BSP checks, on-target smoke/regression tests on a flashed DUT, or both. Not for build or flash steps. Triggers: validate bsp, on-target validation.
atmos-validation
Validate Atmos projects, components, arbitrary JSON Schema inputs, EditorConfig, and GitHub Actions; use affected-file selection and native CI annotations.
skill-benchmark
Benchmark AI skill effectiveness by measuring implementation quality against legacy constraints.