commenting

A guide for posting and replying to comments in Composer threads, which are narrow side panels for discussion. It explains how to choose between a reply, a new comment, and a suggested document change.

In plain words
What is it for?
Use it when replying to a Composer comment, adding a new comment, or deciding whether a proposed edit should be posted as a suggestion.
Why use it?
It keeps comment threads short and readable and prevents posting the same response twice. It also helps put longer structured content in the document instead of the sidebar.

Skill for Claude CodeCodex

Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

agentmods
npx agentmods add skills/hyperactdoting/composer-plugin/commenting
Any agent
npx skills add hyperactdoting/composer-plugin --skill commenting
Clone the repo
git clone --depth 1 https://github.com/hyperactdoting/composer-plugin

Made for: Claude Code, Codex.

Per session 60 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 1,589 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5 $0.00060 $0.01589
Opus 5 $0.00030 $0.00794
Sonnet 5 $0.00012 $0.00318
Haiku 4.5 $0.00006 $0.00159

Measured yesterday against content hash 52c6fe8ae83b, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

commenting scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured yesterday.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

skills/commenting/SKILL.md · 160 lines

How it starts

The opening of the file, as written. The whole thing — 160 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Composer comments and replies

Comment threads in Composer render in a narrow sidebar — think Figma's comment box, not a chat window. Long replies get unwieldy fast. The rules below keep threads readable.

This skill is the full guide for any text reply or new comment. Load it when you're about to call a composer_reply_* or composer_add_comment tool.

Terseness

  • Answer in 1–3 sentences. Prefer one.
  • Reply directly to the question asked — no preamble ("Great question!"), no restating the ask, no trailing summary of what you just did.
  • Never dump your reasoning or tool-call chatter into a comment.
  • If the answer genuinely needs structure (a list of 4+ items, code, a table) and a suggestion isn't the right shape, post it as a suggestion in the doc body instead of as a comment reply. The sidebar is the wrong shape for that content.

Terse beats thorough. If the user wants more, they'll ask.

Don't double-post: a suggestion IS your reply

If you post a suggestion in response to the thread, do not also post a comment reply. The suggestion renders as a Replace/With card in the sidebar already; a pointer comment ("see the suggestion") just duplicates what the user can already see.

Silent suggestion-only responses are correct and expected. The user gets a visible card; the agent's reasoning is in the suggestion's shape, not in a separate sentence.

The only time to post both is when the reply says something the suggestion can't convey on its own — context, a follow-up question, an explicit flag that there are companion edits elsewhere. That's rare; default to suggestion-only.

Reply vs new comment vs suggestion — quick decision

When the user's message in a thread asks for something, pick the shape of your response from this:

Shape Use when
Suggestion only The user asked for a text change ("rewrite this", "make it 200g", "tighten the wording"). The change IS the response.
Suggestion + new comment elsewhere The text change ripples to a different span. Suggest each span separately, then leave a comment on the originating thread saying "also touched X" — only if the ripples aren't otherwise obvious.
Reply only The user asked a question that has a textual answer ("when did we ship that?", "is Jesse still on this?"). No text change involved. Keep it short.
Suggestion + reply on origin thread You changed text AND need to say something the change can't convey alone (caveat, a follow-up question, "I changed two other places too").
New comment elsewhere You're raising something the user didn't ask about but should see — a cross-reference, an inconsistency, a heads-up. Anchor it where it actually belongs, not on the thread you were called from.

Read the full file on GitHub · 160 lines

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. yesterday First seen · 160 lines · 60 tokens per session scan A 52c6fe8ae83b

Subscribe to this mod's changes

commenting is a skill published in the GitHub repository hyperactdoting/composer-plugin (2 stars, last pushed 2mo ago), licensed MIT. It adds 60 tokens to every session and 1,589 once invoked, about $0.0003 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.

Related

Other skills, from other repositories

systematic-debugging

Use when encountering any bug, test failure, or unexpected behavior, before proposing fixes.

obra/superpowers · 21 tokens

next-cache-components-adoption

Turn on Cache Components in a Next.js app and resolve the blocking routes it surfaces. Use when the user wants to enable, adopt, or migrate to Cache Components, flip the cacheComponents flag, work through a flood of blocking-prerender / instant validation errors, run the cache-components-instant-false codemod, or…

vercel/next.js · 95 tokens

babysit-pr

Babysit a GitHub pull request after creation by continuously polling review comments, CI checks/workflow runs, and mergeability state until the PR is merged/closed or user help is required. Diagnose failures, retry likely flaky failures up to 3 times, auto-fix/push branch-related issues when appropriate, and keep…

openai/codex · 114 tokens

imagegen

Generate or edit raster images when the task benefits from AI-created bitmap visuals such as photos, illustrations, textures, sprites, mockups, or transparent-background cutouts. Use when Codex should create a brand-new image, transform an existing image, or derive visual variants from references, and the output…

openai/codex · 113 tokens

cpu-profile-analysis

Analyze V8/Chrome CPU profiles (.cpuprofile) and DevTools trace files (Trace-.json). Use when: profiling performance, investigating slow functions, comparing code paths, finding bottlenecks, analyzing timeToRequest, understanding call trees from sampling profiler data, analyzing layout/paint/rendering, investigating…

microsoft/vscode · 71 tokens

next-cache-components-optimizer

Drive a Next.js route to instant navigation by setting up an agentic loop, under Cache Components / PPR, on initial load (hard navigation) and client-side navigation (soft navigation). Encode the goal as a failing @next/playwright instant() e2e and work it to green, one verified route at a time; the shipped test then…

vercel/next.js · 170 tokens