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 skills add NVIDIA-NeMo/Megatron-Bridge --skill review-prgit clone --depth 1 https://github.com/NVIDIA-NeMo/Megatron-BridgeWrote 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/review-pr)<a href="https://agentmods.dev/skills/nvidia-nemo/megatron-bridge/review-pr"><img src="https://agentmods.dev/badge/skills/nvidia-nemo/megatron-bridge/review-pr/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/review-pr"><img src="https://agentmods.dev/badge/skills/nvidia-nemo/megatron-bridge/review-pr.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 | $0.00064 | $0.02439 |
| Opus 5 | $0.00032 | $0.01220 |
| Sonnet 5 | $0.00013 | $0.00488 |
| Haiku 4.5 | $0.00006 | $0.00244 |
Grade A, and why
review-pr 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 9d 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 — 216 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Review PR
Use this skill to review a change in staged passes. Do not use subagents or tmux unless the user explicitly asks for them.
Ground Rules
- Pull the artifact first: read the PR, commit, diff, log, or files under review before forming conclusions.
- After the artifact is read, load only the relevant repo skills for the touched areas, such as
linting-and-formatting,testing,adding-model-support,build-and-dependency,cicd, or a performance skill. - Preserve user changes. Do not edit files during a review unless the user asks for fixes.
- Do not modify
3rdparty/Megatron-LM/. If the reviewed change touches it, flag that boundary. - Do not run the full test suite. Run only targeted checks, using
uv run python -m pytestor the repo-approved command form. - Prefer findings with concrete file and line evidence. Drop low-confidence or purely stylistic comments unless the user asks for a strict style pass.
Repository Review Principles
Apply these principles to implementation, APIs, recipes, tests, examples, and documentation. They are acceptance criteria, not optional polish.
- Correctness is the gate. Preserve numerical semantics, distributed invariants, checkpoint compatibility, API contracts, and failure behavior. Prefer an explicit error over a silent fallback that can produce incorrect training, conversion, or inference results. Performance or simplicity never justifies behavior that is wrong or cannot be validated.
- Performance is a product requirement. Once correctness is established, protect throughput, latency, memory efficiency, and distributed scaling, especially in training hot paths. Do not accept an avoidable regression for cleaner-looking code or a more convenient abstraction without measurements and an explicit tradeoff. Treat extra synchronization, communication, materialization, copies, allocations, and host overhead as review concerns.
- Design for both users and developers. User-facing behavior should have safe defaults, coherent configuration, actionable errors, and discoverable examples. Developer-facing code should have explicit contracts, clear names and ownership, local reasoning, focused tests, and minimal special cases. Readability includes the public workflow as well as the implementation.
- Keep optimized complexity behind clear boundaries. A fast implementation may be internally specialized, but it should not leak accidental complexity into public APIs or every call site. Isolate the optimized path, document its invariants, and test both its behavior and fallback or error path.
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.
- 9d ago First seen · 216 lines · 64 tokens per session scan A d3b4ea7eddc2
review-pr is a skill published in the GitHub repository NVIDIA-NeMo/Megatron-Bridge (904 stars, last pushed today), licensed Apache-2.0. It adds 64 tokens to every session and 2,439 once invoked, about $0.0003 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
autoreview
Pre-commit/ship code review: Codex default; optional Claude or Pi.
rework-rate
Measure and interpret PR rework rate — the emerging 5th DORA metric.
omh-code-review
This is a Hermes-native code-review workflow skill.
revdiff-plan
Review the last Codex assistant message (plan, analysis, or proposal) with inline annotations in a TUI overlay. Extracts the most recent response from Codex rollout files and opens it in revdiff for review and annotation. Activates on "revdiff-plan", "review plan with revdiff", "annotate plan", "review last response"…
code-reviewer
Code review specialist focused on patterns, bugs, security, and performance.
agent-teams-simplify-and-harden
Implementation + audit loop using parallel agent teams with structured simplify, harden, and document passes. Spawns implementation agents to do the work, then audit agents to find complexity, security gaps, and spec deviations, then loops until code compiles cleanly, all tests pass, and auditors find zero issues or…