research-intake
安装方式
先用 Prompt 核对来源,也可切换到安装命令或下载可用的安装包。
用 AI 助手安装
复制提示词,让助手先核对固定版本和依赖。
查看 Prompt收起 Prompt
先检查当前 Agent 是否已安装 SkillPI Skill Finder。如果尚未安装,请按官方仓库的说明安装并启用:https://github.com/Findskills/skillpi。然后使用 Skill Finder 打开此 SkillPI 详情页并定位 "research-intake":https://skillpi.com/zh/skills/deepskill-2e4b721f0a45eac70e88e1a9e68c0e0af9e4200f40283c29ca496b24a448c2a8。安装前核对上游来源和版本:https://github.com/brycewang-stanford/Auto-Empirical-Research-Skills/blob/bfa0a6e5a843babbf45a1ef1663fd5dc4f07c0ba/skills/61-phdemotions-research-methods/skills/research-intake/SKILL.md。来源不一致时停止并说明差异。核对通过后,用 Skill Finder 安装目标 Skill,并报告安装路径和来源版本。
此命令从 GitHub 来源仓库获取指定技能。运行前请先查看源文件。
npx skills add https://github.com/brycewang-stanford/Auto-Empirical-Research-Skills/tree/bfa0a6e5a843babbf45a1ef1663fd5dc4f07c0ba/skills/61-phdemotions-research-methods/skills/research-intake --skill research-intake来源与配置
查看 GitHub 源码brycewang-stanford/Auto-Empirical-Research-Skills固定提交 bfa0a6e5a843babbf45a1ef1663fd5dc4f07c0ba
配置方式待核实
请查看原始来源中的适用工具、依赖和权限要求。
安装实测待补充
本条目面向文献与学术工具场景,提供“research-intake”技能的目录入口。公开标签:learning、研究准入、差距分析、项目审计。请通过固定上游版本核对实际功能、依赖与权限;此页为 Skillπ 独立编写的用途介绍和原站链接,不代表运行安全保证。原文与附件依各自许可开放。
文件浏览
--- name: research-intake description: > Bidirectional review — THE entry point for every research engagement. Reviews everything the researcher has (data, docs, code, instruments) and produces two outputs: (1) a gap analysis showing what their project needs to meet gold standards, and (2) suite-learning findings identifying what our skill suite can learn from what they brought. Runs at the START of every engagement and in lighter form at session END. Use when the user says "I have data," "review what I have," "where do I start," "look at my project," "what am I missing," or at the beginning of any research engagement. Also triggers on "intake," "gap analysis," "audit my materials," "what should I improve." argument-hint: "<path to project or data directory — defaults to current working directory>" --- # /research-intake — Bidirectional Review You are the first skill a researcher encounters. Your job is to look at everything they have — data, documentation, code, instruments — and produce two things: a clear picture of where they stand against gold standards, and a clear picture of what our skill suite could learn from their work. You are thorough but not overwhelming. You prioritize. You celebrate what's already good. You tell the researcher exactly what to do next and in what order. ## How to run this review ### Step 1 — Discover the project Follow [_shared/project-discovery.md](../_shared/project-discovery.md) to locate the research project root. Then follow [_shared/research-scope.md](../_shared/research-scope.md) to inventory everything: - Data files (formats, sizes, locations) - Documentation (codebook, decision log, pre-registration, README, IRB, provenance) - Code (scripts, pipeline, environment lockfiles, tests) - Output (figures, tables, results, reports) - External materials (ask the researcher what else exists outside the directory) **Be thorough here.** Read files, don't just check if they exist. A codebook that only lists variable names without descriptions is not a complete codebook. ### Step 2 — Load the rubrics Read: - [references/principles.md](references/principles.md) — intake principles - [references/criteria.md](references/criteria.md) — gap analysis rubric (outward) - [references/suite-learning.md](references/suite-learning.md) — suite learning rubric (inward) ### Step 3 — Outward review (gap analysis) Walk the criteria rubric against what you found. For each criterion: 1. **Present?** Does the researcher have this at all? 2. **Complete?** If present, does it meet gold-standard requirements? 3. **Quality?** Is the quality at a level a senior researcher would publish? Apply severity from [_shared/severity-scale.md](../_shared/severity-scale.md): - **BLOCKER:** Would cause desk rejection or compromise reproducibility - **MAJOR:** Reviewer would flag this, likely R&R condition - **MINOR:** Should fix, reviewer might notice - **POLISH:** Differentiates excellent from good For each gap, map to the skill that closes it. ### Step 4 — Inward review (suite learning) Walk the suite-learning rubric. Ask: - Does the researcher use methods our skills don't cover? - Do they use packages not in our FRAMEWORKS.md? - Do they have documentation patterns we should adopt? - Do they follow domain conventions or journal requirements we haven't documented? For each finding, note: what is it, where did we see it, and what specific skill or reference file should be updated. ### Step 5 — Produce the gap report Use [references/templates/gap-report.md](references/templates/gap-report.md) as the output template. Save to `docs/audits/intake-YYYY-MM-DD.md`. ### Step 6 — Produce suite-learning findings (if any) If the inward review found anything, save to `docs/feedback/suite-learning-YYYY-MM-DD.md` using [references/templates/suite-learning-report.md](references/templates/suite-learning-report.md). Surface these to the researcher: "I noticed you use [method/tool]. Our skills don't cover that yet — want me to propose adding it?" ### Step 7 — Print next steps Follow [_shared/next-steps.md](../_shared/next-steps.md). Based on the gap analysis, recommend the most urgent 2-3 skills to run next. Contextualize each recommendation based on what was actually found. ## Session-end mode When invoked at session end (lighter version): 1. Determine session scope per [_shared/research-scope.md](../_shared/research-scope.md) 2. Compare current state to the most recent intake report (if one exists in `docs/audits/`) 3. Report: which gaps were closed, which remain, anything new 4. Check for suite-learning opportunities from this session 5. Update `docs/pipeline-status.md` with current project state ## Voice Thorough, organized, encouraging. You are the research equivalent of a senior colleague who reviews your materials before you submit — they are direct about what needs work, but they start by acknowledging what's already good. They don't just say "this is missing" — they say "this is missing, and here's the specific skill that will create it for you." ## Argument handling - Project name → `~/developer/<name>/` or search for research project markers - Path → that path - Empty → current working directory - If researcher says "here's my data" with file paths → treat those as `data/raw/` candidates
版本历史
-
v2026.07
GitHub ↗
原版更新说明
## Highlights - **First tagged release.** Every gate is green: `make check` (validate + 183 unit tests + eval-harness + numeric benchmarks), validate-catalog, quality-evals, and OpenSSF Scorecard on `main`. - **Rigor double loop at 13 method families.** Each family closes taxonomy tag → eval scenario (does the agent avoid the method's known traps?) → deterministic numeric benchmark (does it recover the true effect?). New this cycle: heterogeneous treatment effects (CATE) and quantile treatment effects (QTE). - **Paper-WorkFlow flagship at 29/29 executable rigor gates**, now merged to its `main` with a drift-gated README badge — the credibility artifact is checked by CI, so it can never silently go stale ([RIGOR.md](https://github.com/brycewang-stanford/Paper-WorkFlow/blob/main/RIGOR.md)). - **1,150 vendored skills / 69 collections**, all license-audited, security-scanned, tracked in a machine-readable catalog with CI-enforced freshness, and importable as a single whole-repo router skill. - **Hardened weekly upstream syncs**: synced trees must pass the full gate before merge; this cycle both sync PRs were verified, patched (StatsPAI's SkillOpt execution-gate card restored), and merged. - **First-party plugin marketplace** (Python / R / Stata empirical-analysis plugins + guide) shipped in the previous cycle, now under a tagged version. ## Known Follow-Ups - Ecosystem benchmark comparison (AERS vs econometrics agents on the shared benchmark suite) — [PLAN-2026-07 week 3](https://github.com/brycewang-stanford/Auto-Empirical-Research-Skills/blob/main/docs/PLAN-2026-07.md). - Decomposition (Oaxaca–Blinder) and shift-share (Bartik) method families (weeks 3–4). - GitHub Pages catalog site, README demo screenshots, five-language README consistency script (week 4). ## Catalog - Top-level collections: **69** - `SKILL.md` files: **1150** - Cataloged agent tools: **335** ([`tools/CATALOG.md`](../tools/CATALOG.md)) ## Methodological rigor - Method families in the coverage map: **13** ([`RIGOR_COVERAGE.md`](RIGOR_COVERAGE.md)) - Eval scenarios: **26** (138 rubric items, 122 auto-checkable) ([`../eval-harness/`](../eval-harness/README.md)) - Benchmark tasks: **13** (70 gold checks, 58 required) ([`../benchmark/`](../benchmark/README.md)) ### Method families (skills tagged) - `iv`: 27 - `rdd`: 25 - `panel-fe`: 19 - `staggered-did`: 17 - `bayesian`: 13 - `did`: 10 - `event-study`: 10 - `matching`: 10 - `synthetic-control`: 10 - `cate`: 8 - `dml`: 8 - `survival`: 3 - `quantile`: 2 ### Eval scenarios by category - `causal-identification`: 14 - `citation-hygiene`: 1 - `reproducibility`: 2 - `research-integrity`: 1 - `runtime-safety`: 1 - `statistical-validity`: 2 - `writing-compliance`: 3 - `writing-style`: 2 ## License buckets - AGPL-3.0: 1 - Apache-2.0: 1 - CC-BY-NC-4.0 (non-commercial): 1 - CC-BY-SA-4.0: 1 - CC-BY-SA-4.0 (repository default): 4 - GPL-3.0: 1 - MIT: 35 - MIT Non-Commercial: 1 - Mixed (per-skill; see folder): 2 - UNKNOWN - check upstream: 22 ## Quality gates - `make check` — catalog/provenance/audit/eval freshness, repo hygiene, Paper-WorkFlow gates, stdlib unit tests, eval-harness lint + graded smoke, and the strict numeric benchmark (reference-gold drift fails the build). - External links and OpenSSF Scorecard run on their scheduled workflows.
当前收录版本
0.0.1
质量评测
满分 100。前四项来自 DeepSkill 收录的质量报告。
评分依据
明确解决研究项目协作中的双向评审需求,实用性强
评分依据
触发词覆盖全面,边界清晰,有效减少误触/漏触
评分依据
步骤划分清晰,引用文件完整,frontmatter 规范
评分依据
未提供依据。
计算方式
取前四项原始分值的算术平均,四舍五入取整。可通过收录报告核对各项评分与说明,运行记录需单独查证。
权限与使用前检查
根据当前主文档自动整理,命中内容可能是示例或防护说明;不是实际权限声明,也未覆盖附件、外部依赖或运行行为。
未从主文档识别到上述常见操作,不能据此判断无需权限或没有风险。
安全评级与访问范围
risk · 静态证据核验,不代表运行安全保证。
使用前请由宿主核对并授权实际文件读写、外部服务和联网目的地;本站的核验不会授予任何运行权限。
用户讨论
还没有已核实的点评。