Skills DirectorySkills Directory
SkillsLearnSecurityCategoriesDocsCommunityBlog
Sign InSubmit Skill
Skills Directory

Security-tested agent skills for Claude, coding agents, and AI workflows.

Directory

  • Browse Skills
  • All Skills A–Z
  • Claude Skills
  • Claude Code Skills
  • Agent Skills
  • Categories
  • Submit a Skill

Learn

  • Learn Hub
  • Install Claude Skills
  • Write SKILL.md
  • Skills vs MCP
  • Directories Compared

Security

  • Security
  • Methodology
  • Secure Claude Skills
  • Security Badges

Company

  • About
  • Community
  • Blog
  • API Docs
  • Advertise

2026 Skills Directory. All rights reserved.

Back to skills

Research Manager

ASecurity

Research project skeleton and file lifecycle. Load when initializing a `.kilo/` project, migrating or archiving documents, or deciding where to put a new research note. Defines the four-directory layout (proposal/project/archive/knowledge), two-digit numeric prefixes, the global.md index layer, archive SUMMARY.md rules, and final REPORT.md requirements. It does not run experiments (experiment_manager) or judge whether an idea is worth pursuing (research_progress).

5 stars
0 votes
0 copies
0 views
Added 9/20/2026
researchgogit

Security Analysis

A100/100

Scanned 9/20/2026

Install to Claude Code

$npx -y skills add WissingChen/my_code_config --skill research_manager --agent claude-code

Installs into .claude/skills of the current project.

Are you the author of Research Manager?

Add the live security badge to your README — it updates automatically with every re-scan.

Security grade badge for Research Manager
[![Security: A — Skills Directory](https://www.skillsdirectory.com/api/skills/wissingchen-research-manager/badge)](https://www.skillsdirectory.com/skills/wissingchen-research-manager)

More formats (shields.io, HTML) on the badges page.

Download Zip
Files
SKILL.md
---
name: research_manager
description: Research project skeleton and file lifecycle. Load when initializing a `.kilo/` project, migrating or archiving documents, or deciding where to put a new research note. Defines the four-directory layout (proposal/project/archive/knowledge), two-digit numeric prefixes, the global.md index layer, archive SUMMARY.md rules, and final REPORT.md requirements. It does not run experiments (experiment_manager) or judge whether an idea is worth pursuing (research_progress).
---

## Output Contract

- 先说结论,再给必要依据和下一步。
- 默认短句和常用词;术语只在更准确时用,首次出现直接解释。
- 内部状态、流程和检查表默认不展示;只有影响决定或用户明确要求时才展开。
- 外部事实、论文结论和数字附来源;不确定的直接写"尚未验证"或"我推测",不给每句话机械加事实/猜测标签。
- 一段能说清就不用表格;独立要点用列表;只有横向比较才用表格。
- 不写套话、廉价肯定、重复总结和固定收尾。
- 禁用黑话和自造词(赋能、闭环、抓手、对齐、链路、落地、打磨等),直接说具体那件事。
- 写文件前先经用户确认。
- 只在发现具体的过时或重复内容时才提议清理,不作为固定收尾动作。

# Research Manager — 管目录、状态和归档

轻量骨架:四个目录、数字编号、一层索引。`research_manager` 负责目录搬动、归档命名、状态流转。不跑实验,不产出内容。

## 1. 目录结构和状态

```
.kilo/
├── global.md          # 纯索引(→ 进度 + TODO.md);不写目标和规则
├── TODO.md            # 项目进展唯一来源;三层树(主线→阶段→方案)
├── proposal/          # 评估中、卡住、已通过还没开工的方向
├── project/           # 正在实验分支上跑的方向
├── archive/           # 已结束的方向
├── reports/           # 生成的带日期状态快照
└── knowledge/         # 外部资料(知识笔记)
```

| 状态 | 位置 | 谁负责 | 进出条件 |
|---|---|---|---|
| 评估中 | `proposal/NN-slug/` | `research_progress` 建议 | 现实检查过或不过 |
| 卡住 | `proposal/NN-slug/` | `research_progress` 标记 | 数据、接口、代码或资源没查清;卡住期间不做实验设计 |
| 通过 | `proposal/NN-slug/` | `research_manager` 记录 | 用户说 GO 才开实验 |
| 进行中 | `project/NN-slug/`(在 `exp/NN-slug` 分支上) | `experiment_manager` 操作 | 继续、转向、证伪或验证通过 |
| 否决 / 证伪 / 验证通过 / 被取代 / 放弃 | `archive/YYYY-MM-DD-NN-slug/` | `research_manager` 执行 | 终版报告是主要交付物 |

## 2. 命名

- 方向单位:`NN-slug/`(两位数字前缀,小写连字符)。搬动时 ID 不变。
- 例外:`global.md`、`TODO.md`、`00-overview.md`、`SUMMARY.md`、`REPORT.md` 和带日期的归档目录不用 `NN-slug`。
- 归档目录:`YYYY-MM-DD-NN-slug/`。
- 知识笔记:`作者-年份-标题.md` 或 `来源-年份-主题.md`;检索日志是 `knowledge/papers/00-query-log.md`(归 `knowledge_keeper` 管)。

## 3. global.md 和 TODO.md(索引 + 待办)

`global.md` 是纯索引,由 agent 维护,≤60 行:

- 目录索引(proposal、project、archive、knowledge、外部索引)
- 指向 `TODO.md` 和当前项目进度的指针

这里不写目标、规则、解释——那些由项目所有者写在 `AGENTS.md`(见 §4)。

`TODO.md`(同目录)是项目进展的唯一来源,由 `research_manager` 创建和维护,其他技能只读。三层树:

1. 主线:一句话中心问题(锁定,只有用户能改)
2. 大阶段:来自 proposal 的确定阶段(如 SFR / TTO / RL),proposal 状态变化时才更新
3. 实现方案:每个阶段的当前做法(不确定),随实验证据增删换,每次改动在回复里说明原因

方向状态一有变化(阶段推进、方案更换、主线调整)就更新。主线和大阶段后面带指向方向文件的链接(如 `project/03-physcene3d/00-overview.md`),有就指,没有不指;也可链接模块级 TODO(如 `tto_pp/TODO.md`),不复制内容。≤150 行。

## 4. 概览文件和谁说了算

文件打架时的优先级:`AGENTS.md`(所有者写的目标和流程约定)→ `global.md`(索引/指针)→ 方向 `00-overview.md`(当前状态)→ 登记的实验计划(冻结的预期)→ 运行报告(观测证据)→ 知识笔记(外部证据)→ 生成的状态报告(一次性视图,永远不算数)。

- **方向级**:每个 `proposal/NN-slug/` 和 `project/NN-slug/` 都有一个 `00-overview.md` 当入口。只记**当前状态**——状态、主要问题、主要假设、主要矛盾、当前证据、最大不确定性、下一步决定、暂存的问题(不自动激活)、被推翻的假设。历史写在运行报告里,不往这里追加。
- **集合级**:一个目录超过五个条目后,才建或更新它的 `00-overview.md`。
- 项目根的 `AGENTS.md` 由项目所有者手写,agent 严格只读。内容是项目目标、不做什么、成功标准、工具链、关键约束、决策约定——绝不放实验结果、论文笔记、进行中的讨论。agent 最多在被要求时把 `references/AGENTS_template.md` 复制为项目根的 `AGENTS.md`,之后全归所有者维护。

## 5. 归档规则和终版报告

结束一个方向只能经 `experiment_manager` 或用户明确授权。结束一个进行中的方向,必须留下一份自足的带图 `REPORT.md` 作为主要交付物。

`SUMMARY.md` 只是短的结果/索引指针,不复制报告内容。

五种结局:

- **否决**:没进实验就失败
- **证伪**:登记过的预期被稳定推翻
- **验证通过**:已转正,证据保留
- **被取代**:核心问题或假设实质改变
- **放弃**:非科学原因(资源、优先级等)

没进过实验的否决/卡住方向由 `research_manager` 直接归档——没有实验分支,提案里的评审记录(`experiment-plan.md`)就是最终文档,不要求 `REPORT.md`。五种结局的完整走查:`references/lifecycle-trace.md`。

## 6. 状态汇报

被请求时("汇报"、"status report")生成带日期的快照 `.kilo/reports/YYYY-MM-DD-status.md`。状态文档只生成、不手维护;相邻快照做 diff 展示变化。

必备内容:

- 头条:≤3 行概括本期
- 在评方向:每个评估中/卡住/通过的方向,状态加一句话(来自 `proposal/NN-slug/00-overview.md`)
- 进行中的进展:每个方向最近一次 `ENN` 的判定、离停手/做成标准的距离——数据由 `experiment_manager` 提供(运行报告在 `exp/NN-slug` 分支上)
- 本期结束的方向:结局和教训(来自 `archive/*/SUMMARY.md`)
- 下一步和风险:待办(`TODO.md`)加判断

每个说法都要能指到已存在的来源。生成的报告不受行数预算限制。

## 7. 行数预算和清理

人维护的运营文档有硬上限:

- `global.md` ≤60 行;`TODO.md` ≤150 行
- 任何单个运营用 SKILL.md 或方向文档 ≤150 行

豁免:论文手稿、生成的报告/数据、代码、参考文献列表、拆不开的表格。

超预算时:把细节拆到旁边的引用文件,使用点留一行指针;不许靠悄悄删规则来压缩。

只在发现具体的过时或重复内容时才提议归档/删除/合并。过时:描述的状态已不再成立、且被更高优先级的文件取代。重复:同一事实在两个文件里维护——保留归口技能负责的那份,另一份换成指针。原始科学证据保留,除非它的保留规则允许删。

## 8. 读文件的规矩

会话开始时先读 `AGENTS.md`(有的话)和 `TODO.md`,再顺着 TODO 里的链接读对应方向的文件。`global.md` 只是索引备查。不整树加载 `.kilo/`。

## 9. 找哪个技能

请求领域 → 技能 的对照表、长报告任务的串联顺序,单一事实源在技能组总览 `my_skills/00-overview.md` §5,这里不复制。

## References

- `references/templates.md`:本技能管的每个文件的填写模板(`global.md`、`TODO.md`、方向 `00-overview.md`、`SUMMARY.md`、状态快照)
- `references/AGENTS_template.md`:项目根 `AGENTS.md` 的模板
- `references/lifecycle-trace.md`:五种结局的完整走查

Attribution

WissingChenWissingChen
View sourceMore from WissingChen →
SSkills DirectorySkills Directory

Your tool, in front of Claude Code builders.

3 founder slots · $299/mo · GSC-verified traffic · sponsors can never buy grades.

See placements

Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.

Comments (0)

No comments yet. Be the first to comment!

SSkills DirectorySkills Directory

Your tool, in front of Claude Code builders.

3 founder slots · $299/mo · GSC-verified traffic · sponsors can never buy grades.

See placements

Related Skills

Competitor Analysis

This skill provides comprehensive analysis of competitor SEO and GEO strategies, revealing what's working in your market and identifying opportunities to outperform the competition.

1823 votes

Deep Research

Universal deep research agent team. 13-agent pipeline for rigorous academic research on any topic. 7 modes: full research, quick brief, paper review, lit-review, fact-check, Socratic guided research dialogue, and systematic review with optional meta-analysis. Covers research question formulation, Socratic mentoring, methodology design, systematic literature search, source verification, cross-source synthesis, risk of bias assessment, meta-analysis, APA 7.0 report compilation, editorial review...

452202 votes

Paperclip Distill

Use when an operation issue is a Paperclip cursor-window, distill, or backfill — `operationType: "distill"` or `"backfill"` and the body references a Paperclip source bundle for a project or root issue. Turn raw Paperclip activity into a wiki-insightful project page, decisions log, and history note. This skill exists specifically to replace the stiff, datestamp-heavy templated output that the deterministic distiller produces.

805541 votes

Academic Pipeline

Orchestrator for the full academic research pipeline: research -> write -> integrity check -> review -> revise -> re-review -> re-revise -> final integrity check -> finalize. Coordinates deep-research, academic-paper, and academic-paper-reviewer into a seamless 10-stage workflow with mandatory integrity verification, two-stage peer review, and reproducible quality gates. Triggers on: academic pipeline, research to paper, full paper workflow, paper pipeline, end-to-end paper, research-to-publi...

452201 votes

Exa Search

Semantic search, similar content discovery, and structured research using Exa API

304951 votes
View all in research →