<2-3 sentences summarising what the project addresses and solves, the target audience, and the benefit.> This repository offers a plain-language writing skill for AI agents, designed to enhance clarity and conciseness in generated text. It provides a revision view that visually highlights changes, aiding in understanding and refining AI output. The skill targets AI developers and users seeking to improve the readability and effectiveness of AI-generated content.

How It Works

The core of this skill is a set of plain-text rules defined in SKILL.md, meticulously crafted for maximum compatibility with any AI agent capable of processing file inputs. This approach bypasses complex integrations, allowing agents to adopt plain-language writing standards simply by being instructed with these rules, whether via a system prompt or a dedicated skills directory. The skill also incorporates a sophisticated self-revision mechanism. It analyzes its own output, identifies content that does not add value or can be expressed more clearly, and generates an HTML file detailing the modifications. This revision view provides a clear, side-by-side comparison of the first draft, the revised draft, and a diff highlighting specific changes, making the AI's editing process transparent.

Quick Start & Requirements

• Primary install / run command: For Claude Code, clone the repo: git clone https://github.com/shreyashankar/plain-writing-skill ~/.claude/skills/plain-writing. Other agents can use rules from SKILL.md pasted into their instructions.
• Non-default prerequisites and dependencies: None explicitly mentioned beyond an AI agent capable of reading files.
• Estimated setup time or resource footprint: Not specified.
• Links: SKILL.md (rules), assets/revision_template.html (template).

Highlighted Details

• Universal Rule Set: Employs plain markdown (SKILL.md) for writing rules, enabling adoption by any AI agent with file-reading capabilities, promoting broad interoperability and reducing integration friction.
• Automated Revision & Diff: Features an integrated self-checking process that refines AI-generated text for conciseness and clarity by removing redundancy and improving expression.
• Visual Change Tracking: Generates an HTML report using assets/revision_template.html that visually distinguishes removed text (red) and rewritten text (green), with interactive tooltips explaining the rationale behind each edit, offering deep insight into the AI's writing process.
• Cross-Agent Applicability: Designed to be usable not only with specific tools like Claude Code but also with other agents (e.g., Codex, Pi) by simply pasting the rules into their instruction sets.

Maintenance & Community

No specific details regarding contributors, sponsorships, partnerships, deprecations, migrations, or community channels (Discord/Slack) are provided in the README.

Licensing & Compatibility

No license type or compatibility notes for commercial use are specified in the README.

Limitations & Caveats

Effectiveness is contingent on the AI agent's ability to accurately interpret and apply the plain-text rules. Output files are generated in the /tmp directory, which may require specific permissions or configurations. No explicit information on supported platforms or advanced features is provided.