Agent skill
comment-mode
Granular feedback on drafts without rewriting. Generates highlighted HTML with click-to-reveal inline comments. Use when user says "comment on this", "leave comments on", "give feedback on", or asks for feedback on a draft. Supports multiple lenses—editor feedback, POV simulation ("as brian would react"), or focused angles ("word choice only", "weak arguments"). A granular alternative to rewrites that lets users review feedback incrementally without losing their voice.
Install this agent skill to your Project
npx add-skill https://github.com/rohunvora/cool-claude-skills/tree/main/skills/comment-mode
SKILL.md
Comment Mode
Generate highlighted HTML with click-to-reveal comments for granular feedback on text. Alternative to rewrites—user keeps their voice, reviews feedback incrementally.
When to Use
- User says: "comment on this", "leave comments on", "give feedback on"
- User pastes text and asks for feedback/critique
- User wants targeted feedback without a full rewrite
Clarify Before Commenting
Use AskUserQuestion to clarify scope when the request is ambiguous. Avoid over-commenting (overwhelming) or under-commenting (missing the point).
Clarify when:
- No lens specified → ask what angle they want
- Long document → ask if they want full coverage or just key sections
- Unclear audience → ask who the recipient is (affects POV comments)
Skip clarification when:
- Lens is explicit ("comment on word choice only")
- Document is short (<500 words)
- Context is obvious from conversation
Example clarification:
User: "comment on this"
[long doc, no lens specified]
→ Use AskUserQuestion:
"What should I focus on?"
Options:
- "Editor feedback (structure, clarity, word choice)"
- "Recipient POV (how [person] would react)"
- "Specific angle (tell me what)"
Lenses
| Lens | Color | Comment style |
|---|---|---|
| Editor | Yellow (#fff3cd) | observations, suggestions: "weak opener", "add proof here" |
| POV (as person) | Blue (#e3f2fd) | reactions from that person's perspective: "i know this already", "legal would freak out" |
| Focused | Yellow | specific angle only: "word choice", "tone", "weak arguments" |
Comment Style
- Always lowercase
- Short (5-15 words)
- Smart about type: observations when noting, suggestions when alternatives help, reactions when simulating POV
- Only comment where there's something worth saying—sparse beats exhaustive
Output Rules
DO:
- Write to
_private/views/{name}-comments-temp.html - Open with
open _private/views/{name}-comments-temp.html - Use highlight colors matching the lens
- Keep comments sparse and high-signal
DO NOT:
- Comment on everything—only where there's an obvious edit or insight
- Over-explain in comments—keep them punchy
- Mix lenses in one output—pick one and stick to it
Template
<!DOCTYPE html>
<html lang="en">
<head>
<meta charset="UTF-8">
<meta name="viewport" content="width=device-width, initial-scale=1.0">
<title>{title}</title>
<style>
* { box-sizing: border-box; }
body {
font-family: -apple-system, BlinkMacSystemFont, 'Segoe UI', Roboto, sans-serif;
max-width: 700px;
margin: 0 auto;
padding: 40px 20px;
line-height: 1.7;
color: #1a1a1a;
background: #fafafa;
}
h1, h2, h3 { margin-top: 2em; font-weight: 600; }
h1 { font-size: 1.4em; }
h2 { font-size: 1.2em; color: #333; }
p { margin: 1em 0; }
.highlight {
background: {highlight-bg}; /* #fff3cd yellow for editor, #e3f2fd blue for POV */
padding: 1px 4px;
border-radius: 3px;
cursor: pointer;
border-bottom: 2px solid {highlight-border}; /* #ffc107 for editor, #2196f3 for POV */
transition: background 0.15s;
}
.highlight:hover, .highlight.active {
background: {highlight-hover}; /* #ffe69c for editor, #bbdefb for POV */
}
.comment {
display: none;
background: {comment-bg}; /* #1a1a1a for editor, #0052cc for POV */
color: #fff;
padding: 8px 12px;
border-radius: 6px;
font-size: 0.85em;
margin: 8px 0;
line-height: 1.5;
}
.comment.show { display: block; }
.section-break {
border: none;
border-top: 1px solid #ddd;
margin: 2em 0;
}
.legend {
position: fixed;
bottom: 20px;
right: 20px;
background: #fff;
padding: 12px 16px;
border-radius: 8px;
box-shadow: 0 2px 12px rgba(0,0,0,0.1);
font-size: 0.8em;
color: #666;
}
</style>
</head>
<body>
{content with <span class="highlight" data-comment="comment text">highlighted phrase</span>}
<div class="legend">{count} comments · click highlights to reveal</div>
<script>
document.querySelectorAll('.highlight').forEach(el => {
const commentText = el.getAttribute('data-comment');
const commentDiv = document.createElement('div');
commentDiv.className = 'comment';
commentDiv.textContent = commentText;
el.insertAdjacentElement('afterend', commentDiv);
el.addEventListener('click', () => {
const wasActive = el.classList.contains('active');
document.querySelectorAll('.highlight').forEach(h => h.classList.remove('active'));
document.querySelectorAll('.comment').forEach(c => c.classList.remove('show'));
if (!wasActive) {
el.classList.add('active');
commentDiv.classList.add('show');
}
});
});
</script>
</body>
</html>
Color Reference
Editor lens (yellow):
.highlight { background: #fff3cd; border-bottom: 2px solid #ffc107; }
.highlight:hover, .highlight.active { background: #ffe69c; }
.comment { background: #1a1a1a; }
POV lens (blue):
.highlight { background: #e3f2fd; border-bottom: 2px solid #2196f3; }
.highlight:hover, .highlight.active { background: #bbdefb; }
.comment { background: #0052cc; }
Workflow
- User pastes text + asks for comments
- If ambiguous: Use AskUserQuestion to clarify lens/scope
- Read text, identify key phrases worth commenting on (sparse, high-signal)
- Generate HTML with highlights and
data-commentattributes - Write to
_private/views/{name}-comments-temp.html - Run
open _private/views/{name}-comments-temp.html
Examples
Editor lens:
User: "comment on this, focus on word choice"
→ Yellow highlights, suggestions like "vague, try 'specifically'" or "jargon, simplify"
POV lens:
User: "comment on this as brian would react"
→ Blue highlights, reactions like "i already know this" or "legal would push back here"
Focused lens:
User: "leave comments on this, only flag weak arguments"
→ Yellow highlights on weak claims, comments like "needs evidence" or "assumes too much"
Sparse/sharp lens:
User: "comment only where there's an obvious fix"
→ Minimal highlights, only typos, missing words, or clear improvements
Recommended Agent Skills
Expand your agent's capabilities with these related and highly-rated skills.
ai-loading-ux
Design AI loading, thinking, and progress indicator UX. Use when explicitly asked to improve AI waiting states, add thinking indicators, or design loading UX for AI interfaces. Covers reasoning display (chain-of-thought), progress steps, streaming states, and the "elevator mirror effect" for reducing perceived wait time.
arena-cli
CLI tools for Are.na: export blocks, enrich with vision AI, generate views. Use when: (1) exporting Are.na blocks incrementally, (2) enriching images with AI-generated titles/tags/patterns, (3) generating browsable HTML views, (4) searching blocks by UI patterns or tags, (5) visual search results when terminal output is insufficient. Triggers: "export arena", "enrich arena", "sync arena", "arena view", "search arena for [pattern]", "show me [pattern]".
unified-messages
Cross-platform messaging aggregator. Use as DEFAULT when user does NOT specify a platform (telegram/imessage). Provides unified inbox, search, and triage across both platforms. Triggers: "check messages" (no platform), "inbox", "who messaged me", "all my messages", "triage". NOT for platform- specific requests - use tg-ingest for "telegram X" or imsg-ingest for "imessage X".
imsg-ingest
Primary iMessage interface. Full-featured CLI for message export, conversation management, contact resolution, and thread state. Use when user mentions "imessage", "imsg", "text message", "iphone messages", or phone numbers. Triggers: "imessage messages", "export imessage", "text from john", "imessage contacts", "sync imessage". This is STANDALONE - do not defer to unified-messages for iMessage operations.
html-style
Apply opinionated styling to barebones HTML. Use when user has plain/unstyled HTML and wants to apply consistent visual styling. Triggers: style this HTML, apply styling, make this look good, /html-style, or when user shares HTML that needs CSS. Transforms tables, lists, status indicators, buttons, and layouts into a cohesive design system.
mobbin-ux
Research-driven UI/UX improvement using Mobbin. Use when user explicitly requests design pattern research ("use Mobbin", "research design patterns", "find UX patterns for X"). Requires Mobbin account and browser automation (claude-in-chrome MCP). Searches Mobbin for relevant UI patterns, extracts design principles, generates a spec document for approval, then implements. NOT for general UI work—only when user wants research-backed design.
Didn't find tool you were looking for?