Are these comments visible in my final rendered documents?

No, the skill uses HTML comment syntax which remains invisible in standard Markdown viewers like GitHub, Obsidian, and VS Code.

Does this skill work with binary files or images?

No, the Annotations skill is designed specifically for human-readable text and source code files that the agent can read and edit.

Can I use this skill to create and track tasks?

Yes, when Claude creates a task based on an annotation, it will include the specific Task ID in its response for easy tracking.

Which file types are supported by the Annotations skill?

It primarily supports Markdown (.md) and Python (.py), but works with most text-based files including .txt, .toml, and .yaml.

How does Claude distinguish between a question and a direct instruction?

Claude analyzes the comment context; general questions receive a dated response, while direct instructions are executed immediately and summarized in the git commit.

Collaborative Annotations

Name: Collaborative Annotations
Author: nicsuzor

bynicsuzor

0•

コラボレーションツール

Automates asynchronous communication between humans and AI agents through inline HTML comments across Markdown, Python, and text files.

The Annotations skill facilitates seamless human-agent collaboration by scanning, responding to, and managing inline HTML comments within your project files. It allows users to leave persistent notes, questions, or direct instructions (e.g., ) that the Claude agent can identify, answer with dated replies, or convert into actionable tasks. By leveraging universal HTML syntax, these annotations remain invisible in rendered Markdown viewers like GitHub or Obsidian while providing a robust audit trail and communication channel directly within the source code or documentation.

主な機能

01Automated dated responses that preserve context, indentation, and file structure

02Asynchronous human-agent threading using universal HTML comment syntax

030 GitHub stars

04Workflow cleanup to remove resolved threads while archiving context to documentation

05Task management integration that converts instructions into tracked project tasks

06Multi-file scanning to identify pending human queries and instructions across workspaces

ユースケース

01Providing direct refactoring instructions for code files via inline comments

02Executing a 'morning scan' to find and address all pending AI-related questions in a workspace

03Conducting peer-review style feedback loops with Claude on research papers and documentation

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add nicsuzor/aops-dist annotations

For use in Claude.ai and ChatGPT

Download Skill