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 zjunlp/SciAtlas --skill sciatlas-quick-paper-searchgit clone --depth 1 https://github.com/zjunlp/SciAtlasWrote 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/zjunlp/sciatlas/sciatlas-quick-paper-search)<a href="https://agentmods.dev/skills/zjunlp/sciatlas/sciatlas-quick-paper-search"><img src="https://agentmods.dev/badge/skills/zjunlp/sciatlas/sciatlas-quick-paper-search/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/zjunlp/sciatlas/sciatlas-quick-paper-search"><img src="https://agentmods.dev/badge/skills/zjunlp/sciatlas/sciatlas-quick-paper-search.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 | $0.00082 | $0.01263 |
| Opus 5 | $0.00041 | $0.00632 |
| Sonnet 5 | $0.00016 | $0.00253 |
| Haiku 4.5 | $0.00008 | $0.00126 |
Grade A, and why
sciatlas-quick-paper-search 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 10d 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 — 101 lines — stays where its author put it; the contents beside it link to each section on GitHub.
SciAtlas Quick Paper Search
Use this skill to deliver a small, evidence-backed paper search from a blank machine. The only SciAtlas retrieval command allowed in this skill is sciatlas search-papers.
Operating Contract
- Do not call any SciAtlas downstream, author, support, report, config, health, or skill-management command for retrieval. The only SciAtlas retrieval command allowed is
search-papers. - Own the end-to-end novice flow: install or locate the CLI, guide registration, configure environment variables, run retrieval, read artifacts, and write the final answer.
- Run setup commands yourself when tool access is available. Do not ask a novice user to run shell commands unless you are blocked by missing permissions or missing human-only information.
- Ask the user only for values a program cannot know: email address, verification code, returned API token, or a clarification when the topic is genuinely ambiguous.
- Never print the full API token in the final answer. Mask it if you must mention it.
- If any step fails, explain the exact blocker in plain language, fix what you can, and continue until a
search-papersrun succeeds, the final result is produced, or the user must provide a missing human-only value.
Zero-Start Bootstrap
- Check whether the CLI executable exists without invoking a SciAtlas command: use
Get-Command sciatlas -ErrorAction SilentlyContinueon Windows PowerShell orcommand -v sciatlason macOS/Linux. - If
sciatlasis unavailable:- If this repository is present, install locally with
python -m pip install -e ./sciatlas. - Otherwise install from GitHub with
python -m pip install "git+https://github.com/zjunlp/SciAtlas.git#subdirectory=sciatlas".
- If this repository is present, install locally with
- Check whether
SCIATLAS_API_KEYis present in the current shell environment. - If no token is configured, guide the user through registration:
- Open or provide
http://sciatlas.openkg.cn/register. - Ask the user for the email only if the registration page or flow needs it and you cannot fill it yourself.
- Ask for the email verification code when it arrives.
- Ask the user to paste the returned
sciatlas_xxxtoken. The token is shown only once.
- Open or provide
- Configure the current shell before running retrieval:
What ships with it
1 file 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.
- 10d ago First seen · 101 lines · 82 tokens per session scan A fe176f05296c
sciatlas-quick-paper-search is a skill published in the GitHub repository zjunlp/SciAtlas (146 stars, last pushed 1mo ago), licensed MIT. It adds 82 tokens to every session and 1,263 once invoked, about $0.0004 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
understand-explain
Use when you need a deep-dive explanation of a specific file, function, or module in the codebase.
generating-solutions
Assembles four independent techniques for producing better alternatives before evaluation begins: structured brainstorming with its known failure modes, the question burst method for reframing problems through questions rather than answers, BrainSwarming as a silent graph-based alternative to verbal brainstorming…
jurisd-research
Expert Australian/NZ legal research and AGLC4 citation using the jurisd MCP server. Use when finding cases or legislation (AustLII), looking up a provision offline, formatting or resolving citations, building a pinpoint, tracing who-cites-what, or producing an AGLC4 bibliography. Triggers on case law, legislation…
maestro-learn
User-invoked learning toolkit — guided reading, investigation, pattern extraction, or second opinions. Manual /maestro-learn only; NEVER auto-invoke for code exploration or analysis — route those intents to the analyze step via /maestro-next.
twenty-nine-sentence-knowledge-extraction
A structured interview method for turning documents, data, procedures, and expert knowledge into information that can support an ontology. Its 29 question areas cover concepts, rules, processes, permissions, actions, and quality constraints.
weread-digest
A workflow for turning highlights and notes from WeRead, a Chinese digital-reading service, into organized knowledge-base notes and, when requested, social-media drafts. It checks several storage locations before deciding a book is missing notes.