ARIS is a collection of Markdown-based skills that define a workflow for autonomous machine-learning research, including idea discovery, experiment automation, and review loops. Researchers and AI coding agents use it across tools such as Claude Code, Codex, Cursor, and OpenClaw without depending on a single framework. The catalogue entries are ARIS workflow skills and agents.
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/wanshuiyin/Auto-claude-code-research-in-sleepnpx agentmods add skills/wanshuiyin/auto-claude-code-research-in-sleep/render-htmlWrote 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/wanshuiyin/auto-claude-code-research-in-sleep/render-html)<a href="https://agentmods.dev/skills/wanshuiyin/auto-claude-code-research-in-sleep/render-html"><img src="https://agentmods.dev/badge/skills/wanshuiyin/auto-claude-code-research-in-sleep/render-html/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/wanshuiyin/auto-claude-code-research-in-sleep/render-html"><img src="https://agentmods.dev/badge/skills/wanshuiyin/auto-claude-code-research-in-sleep/render-html.svg" alt="Reviewed on agentmods" width="80" height="20"></a>- NVIDIA SkillSpector warn
SkillSpector: 1 finding, up to high
These are SkillSpector’s own severities. On a checked sample its high-severity flags on skills were ~96% false positives — a documented command, a public API, a “never do X” rule — so we show them as a caution to read, not a verdict. Why →
- high Anti-Refusal · line 171 Skill instructs the agent to omit warnings, disclaimers, or ethical commentary. Stripping safety caveats hides risk from the user and is a common jailbreak preamble.Fix: Remove instructions that suppress warnings, disclaimers, or ethical commentary. Let the agent surface safety-relevant caveats to the user.
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 | $0.00092 | $0.05548 |
| Opus 5 | $0.00046 | $0.02774 |
| Sonnet 5 | $0.00018 | $0.01110 |
| Haiku 4.5 | $0.00009 | $0.00555 |
Grade A, and why
render-html 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.
Copies of this mod
1 near-identical copy found in the catalogue:
- render-html — 88% identical, 75 lines differ
How it starts
The opening of the file, as written. The whole thing — 317 lines — stays where its author put it; the contents beside it link to each section on GitHub.
/render-html: Markdown → single-file HTML for human reading
Markdown is for writers. HTML is for readers. ARIS workflow nodes write Markdown (canonical, audit-trail-friendly, machine-parseable).
/render-htmlturns selected artifacts into a polished single-file HTML view for the human who actually has to read them. The Markdown stays the source of truth.
When to use this skill
Use /render-html for ARIS artifacts that have a real human reader:
| Artifact | Why HTML helps | Template |
|---|---|---|
idea-stage/IDEA_REPORT.md |
Ranked ideas + pilot signal + scores feel like a decision dashboard, not a flat list | academic |
review-stage/AUTO_REVIEW.md (+ REVIEW_STATE.json) |
Round-by-round score progression + weakness status; pass --state to embed the JSON |
academic |
paper/KILL_ARGUMENT.md (+ KILL_ARGUMENT.json) |
Per-point attack/defense with <details> Q&A cards + red/yellow/green callouts |
academic |
research-wiki/SUMMARY.md (or research-wiki/index.md) |
Cross-entity cockpit: papers / ideas / experiments / claims at a glance | dashboard |
PAPER_PLAN.md (optional) |
Claims-evidence matrix renders better as a polished table than raw MD | academic |
RESUBMIT_REPORT.{md,json} (optional) |
7-state failure-mode ledger | academic or dashboard |
Do NOT use for:
- LaTeX paper output — the final reader-facing artifact is PDF, not HTML.
SKILL.mdfiles — those are internal LLM-facing protocol..aris/traces/*review traces — forensic debug, not human display.- Every Markdown file in your project — only artifacts that benefit from sticky TOC, callouts, math, or score progressions.
Core invariants
- MD / JSON is canonical, HTML is generated view. Edit the source, then re-render. Do not hand-edit the HTML.
- Cross-model review at the artifact boundary (ARIS invariant). Academic-template HTML — used for the artifacts humans actually read (IDEA_REPORT, AUTO_REVIEW, KILL_ARGUMENT, PAPER_PLAN) — is reviewed by a fresh cross-family Codex thread before being claimed as a finished view. Dashboard-template HTML (cockpit / debug views) skips review by default but accepts
--reviewto force it. See § HTML Review Gate below. - Drift detection. Every rendered HTML embeds the source path, SHA256, and generation timestamp in
<meta>tags AND in the visible page header. If the HTML and source diverge, the meta tells you which version of the source produced it. - Single-file output. No build system, no separate CSS, no
node_modules. Just one.html. - CDN-friendly default,
--offlinefallback. MathJax 3 and highlight.js load fromcdn.jsdelivr.netby default. Pass--offlineto skip both — math will appear as raw$x$, code blocks won't get syntax highlighting, but everything stays readable. - Pure stdlib helper.
render_html.pyuses onlyre,html,hashlib,json,datetime,pathlib,argparse,sys. No pip install required. - Defense-in-depth XSS sanitization. The helper strips
<script>/<style>/<iframe>/<object>/<embed>/<form>/<input>/<button>/<link>/<meta>/<base>tags, allon*event-handler attributes (onclick,onload, …), and rewritesjavascript:/vbscript:/data:href/src/action schemes to#blocked-unsafe-url:. ARIS workflow artifacts should not contain these in the first place, but the sanitizer is the safety net in case an LLM hallucinates one. Markdown text content is HTML-escaped separately and never reaches the sanitizer.
What ships with it
3 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 Changed 2d34c2a87819
- 6d ago First seen · 317 lines · 92 tokens per session scan A 0f9e5b457224
render-html is a skill published in the GitHub repository wanshuiyin/Auto-claude-code-research-in-sleep (15,913 stars, last pushed yesterday), licensed MIT. It adds 92 tokens to every session and 5,548 once invoked, about $0.0005 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 2026-09-03.
Other skills, from other repositories
serving-llms-vllm
Serves LLMs with high throughput using vLLM's PagedAttention and continuous batching. Use when deploying production LLM APIs, optimizing inference latency/throughput, or serving models with limited GPU memory. Supports OpenAI-compatible endpoints, quantization (GPTQ/AWQ/FP8), and tensor parallelism.
modal-serverless-gpu
Serverless GPU cloud platform for running ML workloads. Use when you need on-demand GPU access without infrastructure management, deploying ML models as APIs, or running batch jobs with automatic scaling.
skypilot-multi-cloud-orchestration
Multi-cloud orchestration for ML workloads with automatic cost optimization. Use when you need to run training or batch jobs across multiple clouds, leverage spot instances with auto-recovery, or optimize GPU costs across providers.
serving-llms-vllm
Serves LLMs with high throughput using vLLM's PagedAttention and continuous batching. Use when deploying production LLM APIs, optimizing inference latency/throughput, or serving models with limited GPU memory. Supports OpenAI-compatible endpoints, quantization (GPTQ/AWQ/FP8), and tensor parallelism.
modal-serverless-gpu
Serverless GPU cloud platform for running ML workloads. Use when you need on-demand GPU access without infrastructure management, deploying ML models as APIs, or running batch jobs with automatic scaling.
tensorrt-llm
Optimizes LLM inference with NVIDIA TensorRT for maximum throughput and lowest latency. Use for production deployment on NVIDIA GPUs (A100/H100), when you need 10-100x faster inference than PyTorch, or for serving models with quantization (FP8/INT4), in-flight batching, and multi-GPU scaling.