Doc Coauthoring logo

Doc Coauthoring

Community
Immortalqx
doc-coauthoring

Structured co-authoring workflow for specifications, proposals, decision docs, PRDs, RFCs, and other substantial written artifacts. Use when Codex should help a user gather context, shape document structure, draft section by section, and test whether the document works for a cold reader before handoff.

Overview

PublisherImmortalqx
Repositorymy_codex_skills
Skill namedoc-coauthoring
Stars
56
Forks
0
Bundled files
1
LicenseMIT
Links
  • Markdown instructions

    A SKILL.md file the model loads on demand, so it only costs tokens when a request actually matches.

  • Works with any LLM

    AI skills are plain Markdown, not provider-specific code, so this works with GPT, Claude, Gemini, Grok, or a local model.

  • 1 bundled files

    Scripts, templates, and references the model can read while it works. Files are read-only and never executed.

  • Open source

    Published by Immortalqx on GitHub. Read the source before you install it.

Installation

Install the Doc Coauthoring AI skill in TypingMind to use it with any LLM, or drop it into another agent that reads SKILL.md.

1

Install in TypingMind

TypingMind installs a skill straight from its GitHub folder — it reads SKILL.md, bundles the resource files, and stores the result locally.

  1. Open the app and go to Plugins → Skills.
  2. Choose "Install from GitHub".
  3. Paste the skill folder URL below and confirm.
    https://github.com/Immortalqx/my_codex_skills/tree/main/doc-coauthoring
  4. Enable the skill in any chat where you want it available.
Plugins → Skills → Add skill → From GitHub URL, then paste the folder URL and press Continue.
2

Install in another agent

Any agent that reads the Agent Skills format can use this skill — copy the folder into that agent's skills directory.

Claude Code — .claude/skills
git clone --depth 1 https://github.com/Immortalqx/my_codex_skills.git /tmp/my_codex_skills
mkdir -p .claude/skills
cp -r /tmp/my_codex_skills/doc-coauthoring .claude/skills/doc-coauthoring
Restart Claude Code after copying so it picks up the new skill.

Use it in TypingMind

Enable Doc Coauthoring in any TypingMind chat and the model takes it from there. Its name and description sit in the system prompt, and the moment a request matches, the model loads the full instructions itself — you never invoke it by hand, and it costs no tokens until it is actually used.

The model loads Doc Coauthoring on its own as soon as a request matches it.

Works with any AI model

AI skills are plain Markdown instructions rather than provider-specific code, so Doc Coauthoring is not tied to the model it was written for. Install it once in TypingMind and use it with GPT-5, Claude, Gemini, Grok, DeepSeek, Mistral, Llama, or a local model you run yourself — all on your own API keys.

  • Loaded only when it is needed

    The system prompt carries just the name and description. The instructions are fetched on the first matching request, so an idle skill costs nothing.

  • Switch models mid-chat

    Because the skill is instructions rather than code, changing model does not break it — the next model reads the same SKILL.md.

Skill instructions

This is the SKILL.md content the model loads. Read it before installing — a skill is instructions your model will follow.

Doc Co-Authoring

Overview

Use this skill when the user is writing a substantial document and the bottleneck is not grammar, but structure, missing context, or unclear reader expectations. The workflow is:

  1. Gather context until the problem, audience, and constraints are concrete.
  2. Create a scaffold in a local file or the user-requested output format.
  3. Draft the document section by section.
  4. Run a cold-reader test and fix gaps before declaring the document done.

This skill is format-agnostic. Use plain Markdown by default. If the requested deliverable is specifically a Word document, slide deck, or spreadsheet-backed document, hand off the file-format work to docx, pptx, or xlsx after the structure is stable.

Stage 1: Context Gathering

Goal

Build enough shared context that later drafting steps can focus on tradeoffs and clarity instead of rediscovering basics.

Start with meta-context

Ask for:

  1. Document type.
  2. Primary audience.
  3. Decision or action the document should drive.
  4. Required template or section order.
  5. Hard constraints such as deadline, page budget, compatibility requirements, or political constraints.

Accept shorthand answers. If the user already pasted a draft, infer what you can before asking follow-up questions.

Pull in source material

If the user mentions existing notes, tickets, threads, design docs, or files:

  • Read the local files directly when they are available in the workspace.
  • Use MCP or other connected tools only when they are actually available in the current environment.
  • If a referenced source is not accessible, ask the user to paste the relevant parts instead of pretending the source can be read.

Ask clarifying questions

After the initial dump, ask focused numbered questions about the gaps that would block drafting. Aim for 5 to 10 high-value questions, not a questionnaire for its own sake.

Good examples:

  • What is the concrete failure mode that triggered this document?
  • Which alternatives were considered and rejected?
  • What does the audience already know, and what must be explained from scratch?
  • What decision should a reader be able to make immediately after reading?

Exit condition

Move on when you can explain, in one short paragraph, what the document is for, who it is for, and what the main recommendation or deliverable is.

Stage 2: Structure and Scaffolding

Decide the section set

If the user already has a template, follow it. Otherwise propose a minimal structure that matches the document type.

Examples:

  • Decision doc: context, problem, options, recommendation, risks, rollout.
  • Technical spec: goal, requirements, design, edge cases, rollout, open questions.
  • Proposal: motivation, objectives, plan, budget or resources, risks, ask.

Create the scaffold

Create a local file with section headings and placeholder text. Prefer Markdown unless the user explicitly asked for another format.

Placeholder text should be short and functional, for example:

md
## Recommendation

[To be drafted]

Do not overdesign the scaffold. Its job is to make iteration easy.

Stage 3: Section-by-Section Drafting

For each section, run the same loop.

Step 1: Clarify the section

Ask 3 to 7 questions specific to that section. Keep them concrete and answerable.

Step 2: Brainstorm candidate points

List 5 to 15 candidate points depending on section complexity. Include:

  • facts already provided by the user,
  • implications they may have missed,
  • tradeoffs or objections the reader is likely to raise.

Step 3: Curate

Ask the user what to keep, drop, combine, or reorder. If the user gives freeform feedback, translate it into concrete edits.

Step 4: Draft

Replace the placeholder with real prose. Draft for the target audience, not for the agent. Keep claims tied to the actual context gathered in Stage 1.

Step 5: Refine

Iterate surgically. Prefer editing the existing file over reprinting the whole document in chat. Track the user's preferences in tone, level of detail, and section length, then apply them to later sections.

Quality checks during drafting

After a section stabilizes, ask:

  • Can anything be deleted without losing meaning?
  • Is any paragraph doing more than one job?
  • Does this section assume context that the target reader will not have?

Stage 4: Whole-Document Pass

When most sections are drafted, reread the full document and check:

  • flow across sections,
  • duplicated content,
  • contradictions,
  • generic filler,
  • missing transitions,
  • missing definitions for reader-critical terms.

Surface issues as findings, then edit the file directly.

Stage 5: Cold-Reader Test

Goal

Check whether a fresh reader can answer the obvious questions without relying on hidden context from the drafting conversation.

Produce reader questions

Generate 5 to 10 realistic reader questions such as:

  • What problem is this document solving?
  • What decision is being requested?
  • What changed from the status quo?
  • What are the main risks?
  • What follow-up action is required from me?

Testing options

If a fresh subagent, fresh thread, or equivalent independent reader is available, use it with only the document and the reader questions. Do not leak your diagnosis or intended fixes into that test.

If no such isolation mechanism is available, ask the user to test the document manually in a fresh chat or with a real colleague and report back what was unclear.

Fix loop

If the reader test reveals confusion, go back to the exact section that caused it and fix the missing context, ambiguity, or hidden assumption. Repeat until the main reader questions can be answered correctly from the document alone.

Final Review

Before declaring the document complete:

  1. Confirm the document achieves the intended reader outcome.
  2. Ask the user to verify facts, links, metrics, and names.
  3. Check that every section still earns its place.

Working Rules

  • Be direct and procedural.
  • Prefer local files over large chat-only drafts.
  • Do not fabricate access to external systems or documents.
  • Do not optimize for elegance before the structure is correct.
  • If the user wants to abandon the workflow and draft freeform, switch cleanly without argument.

Bundled files

The model reads these on demand while the skill is loaded. They are exposed as readable files and are never executed.

Frequently asked questions

What does the Doc Coauthoring AI skill do?

Structured co-authoring workflow for specifications, proposals, decision docs, PRDs, RFCs, and other substantial written artifacts. Use when Codex should help a user gather context, shape document structure, draft section by section, and test whether the document works for a cold reader before handoff.

Why use Doc Coauthoring on TypingMind?

Because you install it once and use it with any model. Doc Coauthoring is plain Markdown rather than provider-specific code, so the same skill runs on GPT-5, Claude, Gemini, Grok, or a local model — and you can switch model mid-chat without it breaking. TypingMind runs on your own API keys, so you pay providers directly instead of a per-seat subscription, and your skills and chats stay in your own storage.

How do I install Doc Coauthoring in TypingMind?

Open Plugins → Skills → Install from GitHub in TypingMind and paste https://github.com/Immortalqx/my_codex_skills/tree/main/doc-coauthoring. TypingMind reads its SKILL.md and bundles its files and installs it as a skill you can enable per chat.

Which AI models can use Doc Coauthoring?

Any model you connect in TypingMind. AI skills are plain Markdown instructions rather than provider-specific code, so GPT, Claude, Gemini, Grok, and local models can all load this skill when a request matches it.

How many AI models can I use with Doc Coauthoring?

As many as you like. As long as a model supports skills, you can use Doc Coauthoring with it — GPT, Claude, Gemini, Grok, DeepSeek, Mistral, Llama and more — all on TypingMind with your own API keys.

Is the Doc Coauthoring AI skill free?

Yes. It is published on GitHub by Immortalqx under the MIT license. You only pay your own AI provider for the tokens you use.

What are AI skills?

An AI skill is a reusable instruction bundle that teaches an AI model how to do one specific task. It follows the open Agent Skills format: a SKILL.md file with a name and description, plus any scripts, templates or reference files the model may need. The model reads the instructions only when your request matches the skill, so an installed skill costs nothing until it is used.

How are AI skills different from plugins or MCP servers?

A plugin or MCP server gives a model new tools to call — code that runs somewhere and returns a result. An AI skill gives the model knowledge and process instead: how to approach a task, which steps to follow, what good output looks like. Skills are plain Markdown, so they need no server, no API key and no runtime, and they work with any model.

View all

Set up your own AI workspace now

Get notified about new features and future giveaways by subscribing to our newsletter 👇