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-research-planWrote 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-research-plan)<a href="https://agentmods.dev/skills/light0305/light-skills/light-research-plan"><img src="https://agentmods.dev/badge/skills/light0305/light-skills/light-research-plan/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-research-plan"><img src="https://agentmods.dev/badge/skills/light0305/light-skills/light-research-plan.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 | $0.00334 | $0.09988 |
| Opus 5 | $0.00167 | $0.04994 |
| Sonnet 5 | $0.00067 | $0.01998 |
| Haiku 4.5 | $0.00033 | $0.00999 |
Grade A, and why
light-research-plan 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 — 337 lines — stays where its author put it; the contents beside it link to each section on GitHub.
研究方案与实验设计（research-plan）—— 科研主线 stage 5 · 实验矩阵 + 对照公平/可证伪 critical 门
你是 Light 科研流水线的 DAG 第 5 节点。任务不是"写一份漂亮的研究计划",是把 idea-critique 放行的 idea 拆成 院士会逐行追问、能真跑、能复现的实验矩阵,并守住两条最先被枪毙的红线:对照公平(baseline 不放水,否则提升是 假象)和可证伪(假设能被推翻,否则不是科学是包装)。这两条 = critical 一票否决;消融不隔离贡献、统计欠功效 = warn。
一句话定位:把"一屋子做实验的院士在方案评审时真正死磕的"——实验矩阵四要素齐全(假设→变量→指标→停止条件)
- 对照公平(等量调参预算,Dacrema 2019:优化 vs 未优化的比较无法证明 SOTA)+ 消融干净隔离贡献 + 不确定性/功效匹配设计 (多 seed 可估算法随机性,正式 power 只数独立单位)+ 能证伪 + 可复现全留痕(种子含 cuDNN/PYTHONHASHSEED、环境、版本、划分)—— 落成确定性机读门 + critical findings。深度对标真相源 =
docs/competitors/research-plan.md(10 真同类 skill / 7 repo + 机制锚 + 诚实差距)；真实研究者八步资源闭环 =references/research-plan-resource-map.md。谁产 findings、谁是 critical 门(诚实分工):本技能产对照公平/可证伪 critical findings(producer=research-plan,
plan_gate.py四 gate)——fair_baseline(对照放水→critical)、falsifiable(假设无反证条件→critical)被run_checkpoint --stage 5聚合 → critical fail exit 1;ablation_isolation(消融不隔离)、statistical_power(欠功效)= warn 不阻断(spec §4.2 口径)。特殊位置(回炉落点,不是出发点):research-plan 自身门 fail = 改方案,在 stage 5 内修复(reroute 无
ROUTES[5], 对 stage-5 trigger 给manual是诚实兜底——不跨阶段回炉)。但它是别人回炉的目标:7→5(result-analysis 判结果 不支撑假设)、13→5(review-rebuttal 拒稿·实验质疑)→ 总控reroute建议、passport add-back-edge --to 5落账 → 你重规划。是横切常驻吗? 否。这是按需
/调用的主线节点;file-reading(读 idea/数据卡)/memory-pm(记台账/方案变更)/ consistency/research-ethics(预注册防 p-hacking)全程横切常驻,本技能不重复它们。
何时启动(触发信号)
- idea 已通过 idea-critique,要把它拆成可执行可复现的完整实验方案——主用法。
- 用户要"设计实验 / 消融 / 对比 / 敏感性 / 泛化 / 鲁棒性 / 统计显著性 / 算多少种子 / 算算力预算 / 复现某论文"——任一即启动。
- 怀疑"baseline 放水了 / 这假设怎么证伪 / 提升是不是单跑运气"——正中本技能 critical 门。
- 回炉(来自下游):result-analysis 判结果不支撑假设(7→5)、review-rebuttal 拒稿·实验(13→5)→ 带"哪条假设 没撑住 + 效应量/CI"或"审稿人实验质疑原文"重规划——这是决策点,停下问用户(回炉/带病推进/转已知局限)。
你怎么工作:ACT / ASK / NEVER
What ships with it
17 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.
- examples/plan_spec.example.json 2.6 KB
- references.md 30 KB
- references/research-plan-resource-map.md 11 KB
- scripts/failure_tree_gate.py 15 KB runs code
- scripts/plan_gate.py 29 KB runs code
- scripts/plan_lint.py 23 KB runs code
- scripts/power_check.py 14 KB runs code
- scripts/research_package_gate.py 29 KB runs code
- scripts/target_chain.py 17 KB runs code
- templates/experiment_matrix.md 8.6 KB
- templates/failure-tree.example.json 1.9 KB
- templates/plan_package.manifest.example.json 1.0 KB
- templates/preregistration.md 4.3 KB
- templates/reproducibility-checklist.md 3.7 KB
- templates/reproduction-log.md 2.2 KB
- templates/research-plan.md 7.5 KB
- templates/target-chain.example.json 1001 B
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 · 337 lines · 334 tokens per session scan A ab2beecb211a
light-research-plan is a skill published in the GitHub repository Light0305/Light-skills (617 stars, last pushed 2mo ago), licensed MIT. It adds 334 tokens to every session and 9,988 once invoked, about $0.0017 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
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…