AI Collab Playbook skill
AI Collab Playbook bundles 19 skills for Claude Code and Codex: paper review, Codex session recovery, anti-AI-tone editing and research-workflow helpers
- 456 ★ on the repository
- Last commit
Minimal access
Why this level
- The skills are text instructions and templates; they run nothing and make no network calls on their own
Install
- Works in
In your terminal, with SkillFoxx CLI
npx skillfoxx add skills/ai-collab-playbookDetects the agents on your machine, shows the access level and pins the version.
Other ways to install
Access and checks
How we reviewMinimal accessThe tool runs nothing and makes no network calls.
- The skills are text instructions and templates; they run nothing and make no network calls on their own
This is third-party code. Review the repository files before installing.
What it does
This repository by a Chinese PhD student in AI collects a methodology article on working with agents plus 19 working skills under skills/full, each with its own SKILL.md: academic-paper-helper and paper-review-pipeline for reading research papers, paperreview for peer review, collaborating-with-claude and collaborating-with-codex with per-agent collaboration rules, session-recovery-codex to restore an interrupted Codex session, writing-anti-ai to edit out typical AI-generation tells, prompt-polisher to tighten a request's wording, skill-governance-loop and skills-governance to keep a personal skill set from sprawling, and skill-creator to build new ones. Some skills, such as writing-anti-ai, xhs-note-creator and paperreview, have been split by the author into their own repositories and are maintained there, staying here as a bundled copy for anyone who wants one clone instead of ten. The root-level AGENTS.md and CLAUDE.md files describe the rules the author actually follows when working with agents: context before execution, breaking a task into stages, reviewing the result once it finishes. The prompts folder holds four standalone templates with no SKILL.md wrapper: a prompt optimizer, a concept explainer, a video-timestamp summarizer and a tool that turns a research paper into an HTML page. The main article, dated 2026-10-13, lays out the framework the author uses to decide when a task goes to an agent and when a lighter tool, a text-selection toolbar or a messenger, is enough.
Who it is for. For researchers and graduate students who read and write academic text with Claude Code or Codex and want ready collaboration rules plus editing that removes AI-generation tells.
Good fit when
- You need to read and summarize a research paper quickly: academic-paper-helper and paper-review-pipeline break it into parts
- A Codex session got interrupted after a long conversation and the context needs restoring: session-recovery-codex covers that
- A draft reads like it was written by a model, and it needs editing against specific AI-generation tells through writing-anti-ai
- You want to write down your own Claude Code and Codex collaboration rules in the author's style: AGENTS.md and CLAUDE.md are a working example
- A personal skill set has grown unused entries and you want help deciding what to keep: skill-governance-loop walks through that
Not a fit when
- You need one specific skill at its latest version: writing-anti-ai, paperreview and others are maintained separately and the playbook's copy can lag behind
- The work is not in Chinese and not in a research or humanities field: the notes and repository layout follow the author's own AI-research experience
- You need a one-shot Claude Code plugin install: there is no .claude-plugin here or install command, skills are copied into folders by hand
Example request
Read this research paper with academic-paper-helper and prepare a short review brief using paper-review-pipelineLimitations
The README points straight at the author's separate repositories for the current version of several skills, so the skills/full copies can fall behind the latest edits. The prompts folder's templates are not packaged as SKILL.md and are added manually by pasting the text. The repository's primary language is Chinese, with an English README available, but the article and examples assume the author's own research context and point to a small GitHub Discussions thread for feedback rather than formal support.
How to disable. Delete the copied skill folders from ~/.claude/skills or the project directory. The prompts templates need no uninstall: simply stop pasting them.
FAQ
How do I install skills from AI Collab Playbook in Claude Code?
Copy the needed folder from skills/full of the repository, for example academic-paper-helper, into ~/.claude/skills or the project's .claude/skills. There is no separate installer or plugin in the repository.
How does writing-anti-ai here differ from the author's standalone repository?
This is a bundled copy for anyone who wants every skill in one clone. The author states the current version of writing-anti-ai is maintained in the separate cnfjlhj/writing-anti-ai repository, and the playbook copy can lag behind it.
How does this set recover an interrupted Codex session?
The session-recovery-codex skill in skills/full gathers the previous session's context so work can continue without retelling everything from scratch.
Do I need to read Chinese to use these skills?
The main README is in Chinese, with an English README.en.md available. The SKILL.md files inside skills/full read and run independently of the interface language.
What does skill-governance-loop do?
It helps decide which of a user's personal skills are outdated or overlap with each other, and logs those reviews in the skill-governance-data folder.
Discussion
No questions or reviews yet.
Related
SkillsA collection of open Agent Skills by Huashu for content work: slides, design, data processing, video scripts and text proofreading
A skill that edits English drafts: strips AI boilerplate, keeps the author's voice, and can also just list what it found without rewriting
Humanities Writing Companion walks a humanities paper through 12 modes: from sharpening the research question to a blind read and AI-use disclosure
AI Marketing Suite
AI Marketing Suite for Claude Code
A set of marketing skills for Claude Code: site audit, copywriting, emails, content calendars, competitor analysis and client reports
A plugin for novels and serial fiction: a muse agent leads from idea to draft, while writer, critic, editor and canon-check agents each handle one job
Addy Osmani skill for clearer writing: an interview and draft from your own words, a rewrite of existing text, a review that leaves the file alone
Comparison with similar tools
Entries that share the main search query, and how they differ.
| Entry | Official | Repository stars | Main difference |
|---|---|---|---|
| AI Collab Playbook (this entry) | no | 456 | |
| Academic Humanizer | no | 2,435 | a 19-skill bundle for several stages of research work, not a single academic-style text polish |
| Research Paper Writing | no | 7,721 | covers paper review, Codex session recovery and anti-AI-tone editing, not writing the paper text alone |
README in short
The README opens with a link to the main article docs/phd-ai-collab.md dated October 13, 2026, and states the author's position: AI as a colleague, not a tool, with the person staying the main variable in framing the task and judging the result. It continues with a table of repository contents, a list of ten skills spun off into their own repositories with a short purpose for each, and links to discussions and an errata form.