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/redai-infra/relax/code-reviewnpx skills add redai-infra/Relax --skill code-reviewgit clone --depth 1 https://github.com/redai-infra/RelaxWrote 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/redai-infra/relax/code-review)<a href="https://agentmods.dev/skills/redai-infra/relax/code-review"><img src="https://agentmods.dev/badge/skills/redai-infra/relax/code-review.svg" alt="Measured on agentmods" height="20"></a>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 | $0.00046 | $0.01048 |
| Opus 5 | $0.00023 | $0.00524 |
| Sonnet 5 | $0.00009 | $0.00210 |
| Haiku 4.5 | $0.00005 | $0.00105 |
Grade A, and why
code-review 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 5d 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 — 126 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Code Review Expert
Perform a structured review of the current git changes with focus on SOLID principles, architecture, removal candidates, security risks, and Python-specific issues. Tailored for the Relax project (PyTorch distributed training, Ray actors, ML pipeline code).
Default to review-only output unless the user asks to implement changes.
For project structure, coupling points, and key files, see AGENTS.md.
Severity Levels
| Level | Name | Description | Action |
|---|---|---|---|
| P0 | Critical | Security vulnerability, data loss risk, correctness bug, training corruption | Must block merge |
| P1 | High | Logic error, significant SOLID violation, performance regression, gradient issues | Should fix before merge |
| P2 | Medium | Code smell, maintainability concern, minor SOLID violation, missing type hints | Fix in this PR or create follow-up |
| P3 | Low | Style, naming, documentation, minor suggestion | Optional improvement |
Workflow
1) Preflight Context
- Use
git status -sb,git diff --stat, andgit diffto scope changes. - If needed, use
rgorgrepto find related modules, usages, and contracts. - Identify entry points, ownership boundaries, and critical paths (training loop, loss computation, checkpoint saving).
Edge cases:
- No changes: Ask if they want to review staged changes or a specific commit range.
- Large diff (>500 lines): Summarize by file first, then review in batches.
- Cross-package changes: When multiple
relax/subpackages are modified, verify import compatibility first.
2) SOLID + Architecture Smells
What ships with it
5 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.
- 5d ago First seen · 126 lines · 46 tokens per session scan A b39f2a413530
code-review is a skill published in the GitHub repository redai-infra/Relax (580 stars, last pushed 7d ago), licensed Apache-2.0. It adds 46 tokens to every session and 1,048 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
verl-rl-training
Provides guidance for training LLMs with reinforcement learning using verl (Volcano Engine RL). Use when implementing RLHF, GRPO, PPO, or other RL algorithms for LLM post-training at scale with flexible infrastructure backends.
openrlhf-training
High-performance RLHF framework with Ray+vLLM acceleration. Use for PPO, GRPO, RLOO, DPO training of large models (7B-70B+). Built on Ray, vLLM, ZeRO-3. 2× faster than DeepSpeedChat with distributed architecture and GPU resource sharing.
verl-rl-training
Provides guidance for training LLMs with reinforcement learning using verl (Volcano Engine RL). Use when implementing RLHF, GRPO, PPO, or other RL algorithms for LLM post-training at scale with flexible infrastructure backends.
openrlhf-training
High-performance RLHF framework with Ray+vLLM acceleration. Use for PPO, GRPO, RLOO, DPO training of large models (7B-70B+). Built on Ray, vLLM, ZeRO-3. 2× faster than DeepSpeedChat with distributed architecture and GPU resource sharing.
verl-rl-training
Provides guidance for training LLMs with reinforcement learning using verl (Volcano Engine RL). Use when implementing RLHF, GRPO, PPO, or other RL algorithms for LLM post-training at scale with flexible infrastructure backends.
openrlhf-training
High-performance RLHF framework with Ray+vLLM acceleration. Use for PPO, GRPO, RLOO, DPO training of large models (7B-70B+). Built on Ray, vLLM, ZeRO-3. 2× faster than DeepSpeedChat with distributed architecture and GPU resource sharing.