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.
git clone --depth 1 https://github.com/happyhappy-jun/writing-driven-autoresearchWrote 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/plugins/happyhappy-jun/writing-driven-autoresearch/ralphthon-icml)<a href="https://agentmods.dev/plugins/happyhappy-jun/writing-driven-autoresearch/ralphthon-icml"><img src="https://agentmods.dev/badge/plugins/happyhappy-jun/writing-driven-autoresearch/ralphthon-icml.svg" alt="Measured on agentmods" height="20"></a>Grade A, and why
ralphthon-icml 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 8d 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.
What it actually says
{
"$schema": "https://json.schemastore.org/claude-code-plugin.json",
"name": "ralphthon-icml",
"displayName": "Ralphthon @ICML",
"version": "1.0.0",
"description": "Ralphthon @ICML onboarding, the official VESSL Cloud Cookbook autoresearch recipe on A100, local-first W&B observability, Weave tracing, and world-model ideation skills for Claude Code",
"author": {
"name": "Team Attention",
"url": "https://github.com/team-attention"
},
"homepage": "https://luma.com/hjuo7auc",
"repository": "https://github.com/team-attention/ralphthon-icml",
"license": "MIT",
"keywords": [
"ralphthon",
"icml",
"claude-code",
"auto-research",
"karpathy",
"a100",
"vessl-cloud-cookbook",
"world-model",
"wandb",
"weave",
"experiment-tracking",
"observability",
"vessl",
"vessl-cloud",
"vesslctl",
"gpu-compute",
"hackathon",
"event-ops",
"onboarding",
"qr",
"pop"
]
}
What it installs
The manifest is a name and a version. 8 skills travel with it, and installing the plugin installs all of them — 612 tokens a session between them. Each is measured on its own page, and each can be installed alone.
- Skill vessl-cloud-onboarding A 49 tokens
- Skill wandb-track-experiment A 94 tokens
- Skill weave-add-tracing A 103 tokens
- Skill wandb-project-analyst A 96 tokens
- Skill auto-research A 48 tokens
- Skill world-model-ideation A 85 tokens
- Skill hello-ralphthon-icml A 97 tokens
- Skill wandb-onboarding C 40 tokens
What ships with it
1 file beside plugin.json 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.
- 8d ago First seen · 38 lines scan A 9130bd70f6a2
ralphthon-icml is a plugin published in the GitHub repository happyhappy-jun/writing-driven-autoresearch (21 stars, last pushed 1mo ago), licensed Apache-2.0. Its token cost is not measured: this kind of file is read by the harness, not the model. 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 plugins, from other repositories
catalysis-scientist
Reasons from active-site structure, turnover frequency, selectivity, and the Sabatier principle through CO/H2 chemisorption site-counting, Weisz-Prater and Mears transport checks, Langmuir-Hinshelwood/Mars-van Krevelen kinetics, and operando DRIFTS/XAS while treating diffusion-limited apparent rates, DRIFTS spectator…
antimicrobial-resistance-scientist
Reasons from MIC distributions, ECOFFs, and phenotype–genotype concordance through EUCAST/CLSI AST, CARD/RGI and AMRFinderPlus resistomes, GLASS/WHONET/NARMS surveillance, carbapenemase/ESBL/MRSA mechanism panels, and heteroresistance PAP while treating breakpoint mixing, genotype-without-phenotype overclaim…
hypha
Hypha is the AI agent of the Mycelium ecosystem: a connecting filament between engineers and the SysML v2 specification. Provides skills and subagents for KerML/SysML v2 metamodel lookup, standard model library lookup, normative spec citation, and SysML v2 textual-notation validation.
r2g-skills
RTL-to-GDS-to-Graph skill collection. eda-install detects the machine and installs + verifies the open-source EDA toolchain (with a no-sudo conda path). signoff-loop drives an open-source EDA flow (Yosys + OpenROAD-flow-scripts + KLayout + Magic + Netgen + OpenRCX) from a spec or RTL to signed-off GDSII (DRC, LVS…
ar
ar — autoresearch: a skill-only autonomous experiment loop. Propose → measure → keep winners → repeat, against any numeric objective, with noise-floor gating so nothing is banked on noise. Five skills: /ar:start opens a run (dedicated branch + baseline noise floor), /ar:resume advances it, /ar:status prints read-only…
academic-research
Academic research toolkit for AI PhD researchers — literature survey, paper reviewing, citation management, and more.