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/Light0305/Light-skillsnpx agentmods add skills/light0305/light-skills/light-literature-searchWrote 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/light0305/light-skills/light-literature-search)<a href="https://agentmods.dev/skills/light0305/light-skills/light-literature-search"><img src="https://agentmods.dev/badge/skills/light0305/light-skills/light-literature-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/light0305/light-skills/light-literature-search"><img src="https://agentmods.dev/badge/skills/light0305/light-skills/light-literature-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 | $0.00326 | $0.07621 |
| Opus 5 | $0.00163 | $0.03810 |
| Sonnet 5 | $0.00065 | $0.01524 |
| Haiku 4.5 | $0.00033 | $0.00762 |
Grade A, and why
light-literature-search 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 12d 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.
- API 端点真相源:[`references.md`](references.md)(OpenAlex/Crossref/arXiv/S2/Europe PMC/DOAJ/bioRxiv 逐条 curl 核 + 2026-06-17 实测增量) How it starts
The opening of the file, as written. The whole thing — 289 lines — stays where its author put it; the contents beside it link to each section on GitHub.
文献调研(literature-search)—— 科研主线 stage 1 · 在线检索 → 领域地图
你是 Light 科研流水线的第一个 DAG 节点。任务不是"给 20 篇论文列表",是产一张能直接喂 idea-generation 的领域地图:这个方向怎么演化来的(脉络)、有哪几派各自优劣(方法谱系)、哪些坑没填好 (未解问题),外加最像用户设想的那一篇(喂 idea-critique 撞车预警)。
一句话定位:把"一个 Nature/顶会常客做文献调研时真正需要的深度"——时间分层检索 + 建地图而非 罗列 + 主动信号识别 + 诚实标覆盖度——落成确定性脚本编排 + 机读 findings。深度对标真相源 =
docs/competitors/literature-search.md(Round 2 R1:13 个真·同类 skill 拆表 + 机制锚 + 超越点 + 诚实边界)。是横切常驻吗? 否。这是按需
/调用的主线节点;file-reading(读用户给的论文/模板)、memory-pm (记检索式/已读库)、research-ethics/consistency(守门)全程横切常驻,本技能不重复它们。
何时启动(触发信号)
- 用户说"调研一下 X""X 有哪些工作""X 研究现状/前沿""写 X 的 related work""X 方向能不能做""帮我看看这 idea 有没有人做过"——任一即启动。
- 作为流水线第 1 步:在 idea-generation 之前跑,给 idea 喂领域地图 + 撞车预警基线。
- 作为定期追踪:长期项目盯方向新文献(
tracker.py/--from-date增量重跑)。
你怎么工作:ACT / ASK / NEVER
每个动作先归类:该自己做(ACT)、该停下问用户(ASK)、还是绝不(NEVER)?
ACT — 跑确定性检索编排,自己做(不烦用户)
- 三层分别检索分别排序(本技能灵魂,见下「指令流 ①」):
domain_map.py一键出前沿/经典/跨领域三层- 领域地图三件套 + 信号 + (给 idea 则)撞车候选 + 机读 findings。
- 宽 query 治跑题:宽主题必加
--require-terms/--exclude-terms(纯被引排序会把蹭词的领域外 高被引文顶上来——实测搜 "sheep lameness" 不过滤会顶出高被引奶牛/通用文)。 - 滚雪球建脉络:有种子文 →
snowball.py前向(被引)+ 后向(参考)追,补关键词检索盲区。 - 跨领域嫁接:窄领域近三年文稀 →
cross_domain_search.py应用轴×方法轴正交检索(不拼词)找可 迁移的前沿方法。 - 生医方向:
biomedical_search.py(Europe PMC + PubMed MeSH 检索式透传);系统综述 →prisma_flow.py计数勾稽。 - 冻结检索协议:快速/系统性检索先填
templates/search-protocol.example.json，再用search_protocol_gate.py --as-of YYYY-MM-DD核 question/eligibility、至少两类受控独立来源、query ledger、 原始响应 locator+哈希、known-item recall、included seed 引文扩展、筛选计数、覆盖总结、停止证据 locator+哈希、 修订账与非未来日期；并把用户提供/确认的范围、时间窗、语种、文献类型和 review type 作为scope_decision绑定原话/决定记录 locator+hash。FROZEN/REGISTERED/AMENDED必须有协议哈希与冻结日期。 - 检索期防幻觉:可疑 DOI →
verify_citations.py核真实存在(投稿终审交 light-citation)。 - 产 findings:
domain_map.py --report出light.findings.v1(覆盖度 + 撞车),交总控run_checkpoint --stage聚合。
What ships with it
21 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.
- assets/cn_core_issn.csv 1017 B
- assets/litreview_template.md 2.2 KB
- assets/method_card.md 1.5 KB
- examples/goat_littable.md 2.9 KB
- examples/goat_search.json 21 KB
- examples/worked_example_dairy_goat.md 3.2 KB
- references.md 51 KB
- scripts/arxiv_search.py 17 KB runs code
- scripts/biomedical_search.py 18 KB runs code
- scripts/cn_journal_probe.py 7.8 KB runs code
- scripts/cross_domain_search.py 8.4 KB runs code
- scripts/domain_map.py 36 KB runs code
- scripts/pipeline.py 9.2 KB runs code
- scripts/prisma_flow.py 6.3 KB runs code
- scripts/search_normalize.py 46 KB runs code
- scripts/search_protocol_gate.py 32 KB runs code
- scripts/snowball.py 23 KB runs code
- scripts/tracker.py 8.9 KB runs code
- scripts/verify_citations.py 17 KB runs code
- search-resource-map.md 12 KB
- templates/search-protocol.example.json 2.8 KB
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.
- 12d ago First seen · 289 lines · 326 tokens per session scan A b7793a776410
light-literature-search is a skill published in the GitHub repository Light0305/Light-skills (620 stars, last pushed 2mo ago), licensed MIT. It adds 326 tokens to every session and 7,621 once invoked, about $0.0016 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
anti-defensive-writing-en
Stops defensive writing across the entire paper lifecycle — writing, revising, cutting, and organizing experiments. Treats the paper as a press conference, not a project summary, lab log, or self-audit: identify the single most publishable strength of the work and build the most favorable, complete, and persuasive…
anti-defensive-writing
A Chinese-language writing guide for presenting a research paper around its strongest supported contribution. It treats the paper as a focused academic presentation rather than a project diary or complete lab record.
research-writing
A collection of 30 prompt templates for writing and reviewing scientific papers. It covers tasks such as translating, editing, summarizing research, writing sections, creating figure captions, and preparing reviewer replies.
ml-paper-writing
Write publication-ready ML/AI papers for NeurIPS, ICML, ICLR, ACL, AAAI, COLM. Use when drafting papers from research repos, structuring arguments, verifying citations, or preparing camera-ready submissions. Includes LaTeX templates, reviewer guidelines, and citation verification workflows.
ml-paper-writing
Write publication-ready ML/AI papers for NeurIPS, ICML, ICLR, ACL, AAAI, COLM. Use when drafting papers from research repos, conducting literature reviews, finding related work, verifying citations, or preparing camera-ready submissions. Includes LaTeX templates, citation verification workflows, and paper…
academic-citation
Search, verify, and map citations for CS/AI/ML papers. Produces VERIFIED/UNVERIFIED reference lists with Citation-to-Claim maps and Exemplar Sets. Use when: finding references for a paper section, verifying citation accuracy, building exemplar sets for introduction/related work learning, checking if existing citations…