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 ARA-Labs/Agent-Native-Research-Artifact --skill research-visualizergit clone --depth 1 https://github.com/ARA-Labs/Agent-Native-Research-ArtifactWrote 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/ara-labs/agent-native-research-artifact/research-visualizer)<a href="https://agentmods.dev/skills/ara-labs/agent-native-research-artifact/research-visualizer"><img src="https://agentmods.dev/badge/skills/ara-labs/agent-native-research-artifact/research-visualizer/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/ara-labs/agent-native-research-artifact/research-visualizer"><img src="https://agentmods.dev/badge/skills/ara-labs/agent-native-research-artifact/research-visualizer.svg" alt="Reviewed on agentmods" width="80" height="20"></a>- NVIDIA SkillSpector warn
SkillSpector: 1 finding, up to medium
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 →
- medium Excessive Agency · line 199 Skill enables autonomous high-impact decisions without human-in-the-loop verification. Critical operations (destructive commands, financial transactions, data deletion) should require explicit user confirmation.Fix: Add human-in-the-loop confirmation for destructive, irreversible, or high-impact operations. Never auto-execute commands that modify files, send data, or alter system state.
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 | $0.00436 | $0.04551 |
| Opus 5 | $0.00218 | $0.02276 |
| Sonnet 5 | $0.00087 | $0.00910 |
| Haiku 4.5 | $0.00044 | $0.00455 |
Grade A, and why
research-visualizer scanned grade A with 1 finding 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.
Makes network callslowCapability
Not a fault in itself. Listed so you know the mod talks to something, and to what.
allowed-tools: Read, Write, Edit, Glob, Grep, Bash(python3 *|base64 *|find *|ls *|open *|ara *|which *|curl *|lsof *|pkill *|brew *|cargo *|sleep *) How it starts
The opening of the file, as written. The whole thing — 231 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Research Visualizer
You show an ARA. You are a read-only consumer: you read the artifact and emit a view; you never edit the ARA. There are two modes, and one routing decision:
- Export mode (default) — you render the ARA into a single portable HTML file: narrated steps, verbatim evidence, inline figures, and the enrichment overlays. This is the shareable/publishable output. The rest of this document below "What you produce" is this mode.
- Live mode — you drive the official
arabinary (github.com/ARA-Labs/ara-cli):ara checkto validate/lint,ara servefor a local live-reloading viewer with zero LLM calls at view time. Reach for it when the user is mid-edit and wants the view to track saves, wants a validation/CI answer ("does this still pass"), or asks for any of its triggers by name. It renders the ARA's structured fields deterministically but does not (yet) author narrative, inline figure exhibits, or per-node concept/code chips — when the user wants those, or a file they can send to someone, that's export mode. If genuinely unsure which the user wants, ask.
Live mode
- Resolve
<ara-dir>(or, for--hub, an--ara-rootwhose immediate subdirectories are each an ARA, e.g. this repo'sexamples/). which ara→ if missing, followreferences/ara-install.md(Homebrew first, Cargo fallback; never install without the user's confirmation). If present, check the version against the repo's CI pin per the same file — flag an older binary, don't silently upgrade.ara check <dir>first (add--strictto fail on warnings;--jsonfor machine-readable output). Clean → continue. Fixable issues and the user wants them fixed →ara check <dir> --fix, re-check, report what changed — never hand-edit the artifact to satisfy the linter. Notrace/exploration_tree.yamlat all → it's raw research input; route to thecompilerskill first, same as export mode's precondition.ara serve <dir> --port <port>(orara serve --hub --ara-root <dir> --port <port>) in the background; default port 8080, pick another if bound (lsof -i :<port>). Confirm withcurl -s -o /dev/null -w '%{http_code}'→ 200, and read the bound URL from the process's own stdout line rather than assuming.- Report the URL and open it for the user. Edits under
<dir>live-reload (add--pollonly for filesystems where the watcher misses changes). The server keeps running — tell the user how to stop it (e.g.pkill -f "ara serve") so it isn't silently orphaned. If they're new to the viewer, surface the relevant bits ofreferences/ara-serve-ux.md.
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.
- 9d ago First seen · 231 lines · 436 tokens per session scan A 8d6bc306d393
research-visualizer is a skill published in the GitHub repository ARA-Labs/Agent-Native-Research-Artifact (676 stars, last pushed 15d ago), licensed MIT. It adds 436 tokens to every session and 4,551 once invoked, about $0.0022 per session on Opus 5. A static security scan graded it A with 1 finding (makes network calls). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other skills, from other repositories
html-ppt-zhangzara-monochrome
A grant proposal on CRISPR base-editing for sickle-cell disease — the hypothesis, the approach, the milestones, and the risk. Built as a decision-grade academic research deck for grant review committee.
html-ppt-zhangzara-pin-and-paper
A field-biology capstone on urban pollinator decline — the survey design, the data, the contribution, and the caveats. Built as a decision-grade coursework defense deck for faculty reviewers.
paper-illustration
A workflow for generating academic illustrations, such as architecture diagrams and method visuals, with image generation and repeated review. Claude plans and checks the figure during the process.
paper-illustration-image2
Generate publication-quality academic illustrations through a local Codex app-server bridge that uses Codex native image generation. This is a separate experimental alternative to paper-illustration, intended for Claude Code users who want a GPT-image-style renderer without modifying the original skill.
paper2video
Turn a research paper, a paper2assets package, or an existing PPT deck into a narrated MP4 video by fully delegating slide authoring to the installed ppt-master skill and fully delegating rendering, subtitles, timeline assembly, and strict media QA to the installed pptx2video skill and its public CLI. Resolves one…
figure-composer
Compose one publication-grade multi-panel figure. Start from a one-line claim plus immutable data Artifact Version references, or inspect an existing figure and draft its outline directly. Plan a 12-column panel outline, delegate one worker per panel, compose and inspect the result, then run at most three adversarial…